Human FKBP5 Negatively Regulates Transcription through Inhibition of P-TEFb Complex Formation

ABSTRACT

Although a large number of recent studies indicate strong association of FKBP5 (aka FKBP51) functions with various stress-related psychiatric disorders, the overall mechanisms are poorly understood. Beyond a few studies indicating its functions in regulating glucocorticoid receptor, and AKT signaling pathways, other functional roles (if any) are unclear. Here, we report an antiproliferative role of human FKBP5 through negative regulation of expression of proliferation-related genes. Mechanistically, we show that, owing to the same region of interaction on cyclin-dependent kinse 9 (CDK9), human FKBP5 directly competes with cyclin T1 for functional positive transcription elongation factor b (P-TEFb) complex formation. In vitro biochemical assays, coupled with cell-based assays, showed a strong negative effect of FKBP5 on P-TEFb-mediated phosphorylation of diverse substrates. Consistently, FKBP5 knockdown showed enhanced P-TEFb complex formation that led to increased global RNA polymerase II C-terminal domain (CTD) phosphorylation, expression of proliferation-related genes, and subsequent proliferation. Thus, our results show an important role for FKBP5 in negative regulation of P-TEFb functions within mammalian cells.

INTRODUCTION

The human FKBP5 is a 51-kDa protein of the immunophilin family that binds to immunosuppressants such as rapamycin and FK506 (1, 2). The functions of this protein have largely been described in context of the HSP90 cochaperone complex (3). A recent surge in reports indicates a strong association of FKBP5 functions with several neurological diseases, including posttraumatic stress disorder (PTSD) (4–6). Furthermore, FKBP5 functions have also been correlated with multiple other diseases and processes, including type 2 diabetes, adipogenesis, and fatty acid metabolism, as well as cancers (7). In several cancers, a strong negative correlation has been observed between FKBP5 expression and severity of disease (8–11). However, little is known about the molecular mechanisms of function of this protein and its role in pathogenesis of multiple diseases.

The well-known mechanistic functions of FKBP5 have been described in the context of AKT and glucocorticoid (GC) signaling pathways (3, 7, 9, 10). FKBP5 has been shown to regulate the phosphorylation status of one of the key regulatory residues (Ser473) of AKT1 through regulating the recruitment of PH domain leucine-rich repeat phosphatase (PHLPP) (8). FKBP5 has also been shown to negatively regulate GC signaling by reducing GC binding to cognate receptors and thus reducing nuclear translocation and downstream activation of target genes (12–14). Although a plethora of literature suggests a strong connection of FKBP5 functions with neuronal abnormalities, recent studies have also indicated its role in controlling type 2 diabetes, in part through control of AKT signaling and the downstream effector protein AKT substrate 160 (AS160) (15). Besides, numerous other studies have also shown a negative correlation between FKBP5 expression and glucose uptake in the plasma membrane (15–17). Reduced levels of FKBP5 have also been strongly correlated with enhanced expression of UCP-1 and PRDM16 genes, the key regulators involved in conversion of white adipose tissue (WAT) to brown adipose tissue (BAT) during adipogenesis (18). Several aggressive cancers, including breast and prostate cancers, also show reduced expression of FKBP5 (19). All of these observations strongly point toward a negative role of FKBP5 in regulating the expression of multiple target genes. However, the underlying mechanism(s) of this negative regulation is completely unknown.

Among all the positive regulators of transcription, the human positive transcription elongation factor b (P-TEFb) complex plays an important role in activation of target gene expression (20–22). The human P-TEFb complex is a heterodimer composed of the kinase subunit cyclin-dependent kinse 9 (CDK9) and associated cyclin T1/T2 (23, 24). The majority of the human P-TEFb complex remains in the inactive 7SK small nuclear RNA (snRNA) complex in association with HEXIM1, LARP7, MePCE2, and 7SK snRNA (25, 26). During transcriptional activation, upon exposure to external cues such as stress, the P-TEFb complex dissociates from the inactive 7SK snRNA complex to associate with multiple transcriptional activators, including Brd4, p53, ZMYND8, and components of the super elongation complex (SEC) (22, 27–31). Upon recruitment at the target sites by the associated activator proteins at the promoter-proximal region, the P-TEFb complex has been shown to play a major role in releasing paused RNA polymerase II (Pol II) into the productive elongation step through phosphorylation of target NELF and DSIF complexes, as well as conserved Ser2 and Ser5 residues of the C-terminal domain (CTD) of Pol II (32–35). Therefore, controlling the functions of P-TEFb complex is gradually turning out to be a major paradigm for activation of target gene expression and associated multiple cellular processes.

In an effort to understand the functional regulation of FKBP5 in negative regulation of expression of target genes, we were intrigued by our observation of a strong association of FKBP5 and CDK9 within mammalian cells. Detailed mechanistic analysis showed that human FKBP5 directly competes with cyclin T1, thereby regulating the level of P-TEFb complex formation and thus directly inhibiting P-TEFb complex-dependent phosphorylation of multiple substrates, including Ser2 and Ser5 residues of the Pol II CTD. Inhibition of this phosphorylation activity of the P-TEFb complex directly reduces expression of diverse target genes, including proliferation-related genes such as CCND1 and C-MYC. Consistent with this observation, FKBP5 knockdown cells show enhanced expression of these target genes, which results in increased proliferation. Thus, we describe here a novel transcriptional repression role of human FKBP5 through negative regulation of functions of the P-TEFb complex within mammalian cells.

RESULTS

FKBP5 is a novel interactor of CDK9.

The P-TEFb complex has been shown to positively regulate transcription through phosphorylation of NELF and DSIF complexes and the CTD of Pol II (21, 22, 35–37). Multiple other regulators have been described that regulate P-TEFb functions either directly or indirectly. Since P-TEFb plays a major role in activating target gene expression, we initially addressed whether any novel factor, not previously described, would be involved in the functional regulation of P-TEFb within mammalian cells. We purified CDK9-associated protein complex using nuclear extract from a stable cell line expressing FLAG-hemagglutinin (HA)-CDK9 protein (22). Interestingly, and much to our surprise, subsequent mass spectrometric analysis showed the presence of FKBP5, along with that of other known CDK9 interactors (see Fig. S1A in the supplemental material). Subsequent immunoprecipitation and blotting analysis further confirmed CDK9 interaction with FKBP5 along with its known interactors, such as SEC components (Fig. 1A). To address whether a reciprocal interaction would also be observed, we generated another stable cell line expressing FLAG-HA-FKBP5 protein (Fig. S1B). Subsequent purification of FKBP5-associated protein complex showed presence of CDK9 component of the P-TEFb complex only along with its known interactor HSP90 (Fig. 1B). Further immunoprecipitation and blotting analyses confirmed that human FKBP5 associates only with the CDK9 component of the P-TEFb complex and not with other CDK9 interactors, such as SEC components (Fig. 1C).

FIG 1.

Novel interaction between FKBP5 and CDK9. (A) Immunoprecipitation and subsequent Western blotting showing identification of FKBP5 as a novel CDK9-interacting protein along with other known super elongation complex (SEC) components. (B) Purification of FKBP5-associated protein complex from a stable cell line that ectopically expresses FKBP5 as FLAG-hemagglutinin (HA) tagged. Eluted proteins were run on a 4 to 12% gradient gel and silver stained for visualization. Individual bands were marked based on their size and subsequent confirmation by Western blotting. A similar parallel immunoprecipitation from a control cell line shows the absence of these bands. (C) Immunoprecipitation and subsequent Western blotting showing identification of CDK9 as a novel FKBP5-interacting protein along with its known interactor HSP90. SEC components failed to show any interaction in this immunoprecipitation assay. (D) Immunoprecipitation of endogenous FKBP5 using specific antibody and subsequent Western blotting showing FKBP5 interaction with CDK9 along with HSP90 in endogenous context. (E) Immunoprecipitation of endogenous CDK9 using specific antibody and subsequent Western blotting showing CDK9 interaction with FKBP5 along with SEC components in endogenous context. (F) Western blotting showing copresence of FKBP5 and CDK9 in the nuclear compartment of mammalian 293T cells. As an indicator of this localization, histone H3 protein is used as a nuclear localization marker, whereas the actin protein acts as a cytoplasmic localization marker.

To rule out artifacts of protein-protein interactions due to overexpression and to address FKBP5 and CDK9 interaction in an endogenous context, we immunoprecipitated endogenous FKBP5 using specific antibody and confirmed its interaction with CDK9 along with its known interactor HSP90 in 293T cells (Fig. 1D). A reverse immunoprecipitation of endogenous CDK9 by using specific antibody also confirmed its interaction with FKBP5, as well as with SEC components, in mammalian cells (Fig. 1E). Therefore, we conclude that FKBP5 and CDK9 interact with each other within mammalian cells and that this interaction is independent of the P-TEFb complex and its associated interactors, such as SEC components. Interestingly, we have observed predominant presence of both FKBP5 and CDK9 within the nuclear compartment of 293T cells (Fig. 1F). This evidence further indicates a functional regulation of CDK9 by FKBP5 through its presence within the nucleus of mammalian cells.

The interaction of FKBP5 with CDK9 is direct and specific.

In order to address whether FKBP5 directly interacts with CDK9 or if any other mediator is responsible for this association, we initially checked their interaction through their expression in heterologous Sf9 cells. We coinfected Sf9 cells with baculoviruses expressing FLAG-FKBP5 and cyclin T1 or CDK9 or both. Subsequent immunoprecipitation analysis confirmed that FKBP5 specifically interacts with CDK9 but not with the cyclin T1 subunit of the P-TEFb complex (Fig. 2A to C). To further address this interaction in vitro, we purified CDK9 and FKBP5 through their expression in mammalian cells and bacterial expression systems, respectively (Fig. S2A and B). Subsequent interaction analysis in vitro using these purified proteins clearly showed that human FKBP5 directly interacts with CDK9 without the requirement of any other mediator (Fig. S2C).

FIG 2.

The TPR1 domain of FKBP5 is important for its interaction with CDK9. (A) Immunoprecipitation and subsequent Western blot analysis showing interaction of FKBP5 and CDK9 through their coexpression in heterologous Sf9 cells. (B) Immunoprecipitation and subsequent Western blot analysis showing failure of cyclin T1 to interact with FKBP5 when they are coexpressed in heterologous Sf9 cells. (C) Immunoprecipitation and subsequent Western blot analysis showing specific interaction of FKBP5 with CDK9 through their coexpression in heterologous Sf9 cells. Even though cyclin T1 is expressed, it fails to show any interaction with FKBP5 in this heterologous expression-based interaction assays. (D) Domain analysis of FKBP5 for its interaction with CDK9. (Upper) Cartoon diagram showing the important domains of FKBP5. (Lower) Immunoblots represent FKBP5 domains (as indicated) and their interaction with CDK9 as observed through their expression in heterologous Sf9 cells. (E) Coimmunoprecipitation analysis showing interaction of ectopically expressed FKBP5 and CDK9 within mammalian 293T cells. (F) Domain analysis of FKBP5 for its interaction with CDK9 within mammalian 293T cells. The immunoblots represent FKBP5 domains (as indicated) and their interaction with CDK9 as observed through their expression in mammalian 293T cells and subsequent immunoprecipitation and blotting analyses.

To further address whether FKBP5 interaction with CDK9 is specific, we tested its interaction with similar kinases, such as CDK7 and CDK8, through their expression in Sf9 cells. We coinfected Sf9 cells with baculoviruses expressing FLAG-CDK7 and FLAG-CDK8 along with FKBP5. Subsequent immunoprecipitation analysis showed that FKBP5 interacts with neither CDK7 nor CDK8 (Fig. S2D and E). Therefore, based on this evidence, we conclude that human FKBP5 and CDK9 directly interact with each other and that the interaction of FKBP5 with CDK9 is very specific.

The TPR1 domain of FKBP5 plays the most important role for its interaction with CDK9 within mammalian cells.

Human FKBP5 is a multidomain protein that includes peptidyl prolyl isomerase 1 (PPIase), peptidyl prolyl isomerase-like (PPIase-like), and tetratricopeptide repeat 1, 2, and 3 (TPR1 to TPR3) domains (Fig. 2D, upper) (38, 39). To identify the domain of FKBP5 that plays an important role in the overall interaction between FKBP5 and CDK9, we coinfected Sf9 cells with baculoviruses expressing different FLAG-FKBP5 fragments along with CDK9 (Fig. 2D). Subsequent immunoprecipitation analyses using FLAG epitope tag showed that while the deletion of TPR3 retained FKBP5 interaction with CDK9 (Fig. 2D, compare lane 4 with lane 6), further deletion of TPR1 and TPR2 abolished this interaction (Fig. 2D, compare lane 6 with lane 7), thus suggesting a role for these domains in the interaction of FKBP5 with CDK9. Consistent with this observation, a C-terminal fragment containing the TPR1 and TPR2 domains (243 to 457, lane 10) fully retained FKBP5 interaction with CDK9. Interestingly, the presence of additional regions within this C-terminal fragment (130 to 242) containing the PPIase-like domain drastically reduced this interaction. This observation thus suggests that while the TPR1 and TPR2 domains are important for FKBP5 interaction with CDK9, presence of the PPIase-like domain inhibits this interaction in this heterologous system.

To further substantiate this interaction within mammalian cells, we initially tested interactions of these two proteins through coimmunoprecipitation analysis. As shown in Fig. 2E, cotransfection and subsequent coimmunoprecipitation analysis clearly showed FKBP5 interaction with CDK9 within mammalian 293T cells (lane 4). Subsequent similar coimmunoprecipitation analysis using different FKBP5 domains (Fig. 2F, same as that of Fig. 2D) further showed that a deletion of the TPR1 domain of FKBP5 completely abolished its interaction with CDK9 within mammalian cells (Fig. 2F, compare lane 7 with lane 8). The deletion of TPR2 and TPR3 had minimal effect on this interaction. Interestingly, unlike our observation within Sf9 cells, a deletion of the PPIase-like domain had minimal effect on FKBP5 interaction with CDK9 within mammalian cells. The overall differences in cell types used and in the experimental setup could explain this discrepancy in our observation. Nevertheless, a deletion of the TPR1 domain completely abolishes FKBP5 interaction with CDK9 and clearly suggests an important role of this domain in overall interaction between these two proteins within mammalian cells.

FKBP5 negatively regulates the phosphorylation activity of P-TEFb in vitro.

Since FKBP5 shows a strong and direct interaction with CDK9 both in vitro and in vivo within mammalian cells, we subsequently addressed the role of FKBP5 in regulating functions of the P-TEFb complex. The human P-TEFb complex has been shown to phosphorylate the NELF and DSIF complexes, as well as the conserved serine residues present within the heptad repeat (YSPTSPS) sequences of the CTD of Pol II (21, 32–35). To address the direct role of FKBP5 in regulation of overall P-TEFb complex-mediated phosphorylation, we employed in vitro kinase assays. Human DSIF and NELF complexes were purified through their expression in bacterial and baculoviral expression systems (Fig. S3A and B) as described previously (34, 40), whereas the recombinant glutathione S-transferase (GST)–CTD protein was purified through its expression in bacterial system (Fig. S3C) (22, 36). Human P-TEFb complex was purified through expression of its subunits in baculovirus-mediated expression in Sf9 cells (Fig. S3D) (22, 36). Our initial analyses showed that the purified P-TEFb complex efficiently phosphorylates its cognate GST-CTD substrate (Fig. S3E), NELF-A and NELF-E subunits (within the NELF complex), and the SPT5 subunit (within the DSIF complex) (Fig. S3F).

Next, we addressed the role of FKBP5 in P-TEFb-mediated phosphorylation of various substrates. Initial analyses showed that addition of purified recombinant FKBP5 to the reaction mixture strongly inhibited P-TEFb-mediated phosphorylation of both the Ser2 and Ser5 residues of Pol II CTD in a dose-dependent manner (Fig. 3A, compare lane 3 versus lanes 4 to 7). It is interesting to note that the dominant function of P-TEFb, Ser2 phosphorylation, is inhibited more strongly than Ser5 phosphorylation in our assays (Fig. 3A, compare the immunoblots using antibodies against Ser2P versus against Ser5P). To further address whether this inhibition of FKBP5 toward phosphorylation by the P-TEFb complex is substrate specific, we performed similar analyses using purified NELF and DSIF complexes. As shown in Fig. 3B and C, similar to its effect on P-TEFb-mediated phosphorylation of GST-CTD, purified recombinant FKBP5 efficiently inhibits P-TEFb-mediated phosphorylation of the NELF and DSIF complexes as well. Therefore, based on all these results, we conclude that human FKBP5 negatively regulates P-TEFb-mediated phosphorylation of diverse substrates in vitro.

FIG 3.

FKBP5 strongly inhibits P-TEFb-mediated phosphorylation of diverse substrates in vitro. (A) In vitro phosphorylation and subsequent Western blot analysis showing inhibition of P-TEFb-mediated phosphorylation of Ser2 and Ser5 residues of the Pol II C-terminal domain (CTD) by FKBP5. For each reaction mixture, 30 μM glutathione S-transferase (GST)–CTD, 15 μM P-TEFb in each lane as appropriate, and 30 μM, 60 μM, 120 μM, and 240 μM (in lanes 4, 5, 6, and 7, respectively) of purified FKBP5 were added. Antibodies specific to phosphorylated Ser2 and Ser5 residues were used in immunoblotting analyses for identifying P-TEFb-mediated phosphorylation of target serine residues present in GST-CTD substrate. (B) In vitro phosphorylation and subsequent autoradiography analysis showing inhibition of P-TEFb-mediated phosphorylation of the Spt5 subunit of the DSIF complex by FKBP5 in a dose-dependent manner. γ-³²P-ATP was used for labeling the P-TEFb-mediated phosphorylation of Spt5 subunit and subsequently autoradiographed for detecting phosphorylation events. For each reaction mixture, 40 μM purified recombinant DSIF, 15 μM P-TEFb in each lane as appropriate, and 30 μM, 60 μM, 120 μM, and 240 μM (in lanes 4, 5, 6, and 7, respectively) of purified FKBP5 were added. As observed, increasing addition of FKBP5 inhibited P-TEFb-mediated phosphorylation of Spt5 subunit in a dose-dependent manner. Along with Spt5 phosphorylation, we also observed a strong effect of FKBP5 on the autophosphorylation activity of the P-TEFb complex, predominantly on its CDK9 subunit. “P” denotes the phosphorylated form of indicated proteins in our assay. (C) In vitro phosphorylation and subsequent autoradiography analysis showing inhibition of P-TEFb-mediated phosphorylation of NELF-A and NELF-E subunits of the NELF complex by FKBP5. γ-³²P-ATP was used for labeling the P-TEFb-mediated phosphorylation of both these subunits and subsequently autoradiographed for detecting phosphorylation events. For each reaction mixture, 30 μM purified recombinant NELF, 15 μM P-TEFb in each lane as appropriate, and 30 μM, 60 μM, 120 μM, and 240 μM (in lanes 4, 5, 6, and 7, respectively) of purified FKBP5 were added. As observed, and in this case as well, increasing addition of FKBP5 inhibited P-TEFb-mediated phosphorylation of NELF-A and NELF-E subunits in a dose-dependent manner. “P” denotes the phosphorylated form of indicated proteins in our assay. (D) Purification of recombinant FKBP5 full-length protein and its mutant derivatives through their expression in a bacterial expression system. Protein bands marked with red-filled dots represent the purified protein bands. Others represent either degradation or nonspecific proteins. (E) In vitro phosphorylation and subsequent Western blot analysis showing the effect of different domains of FKBP5 on inhibition of P-TEFb-mediated phosphorylation of Ser2 and Ser5 residues of Pol II CTD. For each reaction mixture, 30 μM GST-CTD, 15 μM P-TEFb in each lane as appropriate, and 250 ng of purified FKBP5 (full-length and deletion fragments) were added.

To identify the specific domain within FKBP5 required for its overall inhibitory role in P-TEFb-mediated phosphorylation, we purified several FKBP5 fragments through its expression in bacterial expression system (Fig. 3D). Subsequent kinase assays showed that, consistent with our earlier observation, full-length FKBP5 efficiently inhibits P-TEFb-mediated phosphorylation of GST-CTD substrate in this assay as well (Fig. 3E, compare lane 4 versus lane 5). Interestingly, a deletion of the C-terminal 385 to 457 amino acid residues further enhanced FKBP5-mediated inhibition of phosphorylation by the P-TEFb complex, suggesting a role of this domain in antagonizing the overall inhibitory effect. Furthermore, and consistent with a role of the TPR1 domain in FKBP5-CDK9 interaction (Fig. 2F), a deletion of the TPR1 domain that prevents its interaction with CDK9 significantly reduced the inhibitory effect of FKBP5 on P-TEFb-mediated CTD phosphorylation compared to deletions of other domains that show maximal inhibition (Fig. 3E, compare lanes 6 to 8 versus lane 9). Consistently, a C-terminal fragment containing the TPR1 domain showed efficient inhibition of phosphorylation compared to the fragments without this domain (Fig. 3E, compare lanes 9 and 10 versus lanes 11 and 12). Overall, based on all of these data, we conclude that human FKBP5 efficiently inhibits P-TEFb-mediated phosphorylation of cognate GST-CTD substrate in vitro. Furthermore, while the C-terminal domain of FKBP5 (385 to 457 amino acids) negatively regulates this inhibitory function, its TPR1 domain-mediated interaction with CDK9 is important, at least in part, for efficient inhibition of P-TEFb complex-mediated phosphorylation of GST-CTD substrate in vitro.

FKBP5 regulates global Pol II CTD phosphorylation within mammalian cells in an HSP90-independent manner.

Since our in vitro analyses showed a negative role of FKBP5 in P-TEFb-mediated phosphorylation of GST-CTD substrate (Fig. 3), we wanted to address whether similar effects would also be observed within mammalian cells. To address this, we generated short hairpin RNA (shRNA)-mediated stable FKBP5 knockdown 293T cells (Fig. 4A, left, immunoblots; right, mRNA analysis). Subsequent immunoblotting analyses using whole-cell extract clearly showed that upon FKBP5 knockdown, significant increases in global Pol II CTD phosphorylation at both Ser2 and Ser5 residues were observed compared to that in control scramble knockdown cells (Fig. 4B; see Fig. S4A for quantitation). Interestingly, in our analyses, we consistently observed a modest increase in CDK9 level upon FKBP5 knockdown within 293T cells. This could as well be a result of FKBP5 being a member of the HSP90 chaperone complex, which uses CDK9 as a client protein. Furthermore, global levels of Pol II were also increased due to increases in Ser2 and Ser5 phosphorylation that were previously shown to protect it from ubiquitin proteasome-mediated degradation by cognate E3 ligase Def1 (41). Nevertheless, normalization of Pol II levels with the levels of phosphorylated Ser2 and Ser5 form of Pol II further confirmed an overall modest but reproducible increase in global levels of Pol II Ser2 and Ser5 phosphorylation upon FKBP5 knockdown within 293T cells (Fig. S4B). Similar results were also observed in both colon carcinoma (HCT116) (Fig. 4C; see Fig. S4C and D for quantitation) and prostate cancer (PC3) cell lines (Fig. 4D; see Fig. S4E and F for quantitation), thus ruling out an effect of FKBP5 on increasing global Pol II CTD Ser2 and Ser5 phosphorylation in a cell type-specific manner.

FIG 4.

Global increase in Pol II CTD Ser2 and Ser5 phosphorylation upon FKBP5 knockdown in various cell lines. (A) Immunoblot analysis showing short hairpin RNA (shRNA)-mediated stable knockdown of FKBP5 within 293T cells by various target shRNAs (left) and quantitative real-time PCR (qRT-PCR) analysis showing FKBP5 mRNA levels in knockdown cells (right). (B) Immunoblot analysis showing the effect of stable FKBP5 knockdown on global levels of various factors (as indicated) within mammalian 293T cells. (C) Immunoblot analysis showing the effect of stable FKBP5 knockdown on the global levels of various factors (as indicated) within mammalian HCT116 colon carcinoma cells. (D) Immunoblot analysis showing the effect of stable FKBP5 knockdown on global levels of various factors (as indicated) within mammalian PC3 prostate cancer cells. (E) Immunoblot analysis showing the effect of reexpression of full-length and CDK9 interaction-defective FKBP5 proteins in FKBP5 knockdown 293T cells on the global levels of various factors, as indicated. (F) Immunoblot analysis showing the effect of geldanamycin (HSP90 inhibitor) treatment (5 μM final concentration) in FKBP5 knockdown cells on an overall increase in the global level of Pol II CTD phosphorylation in 293T cells. The level of CDK9 was used as a control for geldanamycin treatment. (G) Immunoblot analysis showing the effect of FKBP5 knockdown on interaction between ectopically expressed FLAG-CDK9 and endogenous HSP90 compared to that in control scramble cells.

To further address whether the effect of FKBP5 knockdown on global increases in Pol II Ser2 and Ser5 phosphorylation is specific to FKBP5 only and not an indirect effect, we reexpressed full-length FKBP5 in the knockdown cells. As shown in Fig. 4E, reexpression of full-length FKBP5 reversed the effect of FKBP5 knockdown on those phosphorylation level (Fig. 4E, compare lane 2 versus lane 3). Further, and more importantly, an FKBP5 fragment (1 to 157) that failed to interact with CDK9 (Fig. 2D and F) and show an effect on reducing overall FKBP5-mediated inhibition of CTD phosphorylation in vitro (Fig. 3E) failed to reverse the effect of FKBP5 knockdown on global increases in Ser2 and Ser5 phosphorylation (Fig. 4E, compare lane 3 versus lane 4). Overall, these results further substantiate a negative role of FKBP5 in P-TEFb-mediated phosphorylation both in vitro and in vivo within mammalian cells. Furthermore, these results also signify a role of FKBP5-CDK9 interaction in the overall inhibitory effect of FKBP5 on P-TEFb-mediated phosphorylation of Pol II CTD.

The functions of FKBP5 have been described in the context of a cochaperone of HSP90 (3). Furthermore, CDK9 has also been described as an HSP90 client protein (42). Therefore, it is possible that an FKBP5 knockdown affects the functions of HSP90 toward its client protein such that enhanced folding of the CDK9 protein, leading to enhanced P-TEFb complex formation, may result in increased Pol II Ser2 and Ser5 phosphorylation, as observed in multiple cell lines. To rule out this possibility and show that an overall effect of FKBP5 knockdown is independent of HSP90 functions, we treated the cells with geldanamycin, a known HSP90 inhibitor. As shown in Fig. 4F, although treatment of cells with geldanamycin results in a reduced level of CDK9 (compare lane 2 versus lane 3), a known effect reported previously (43), it failed to show any effect on global increases in Pol II Ser2 and Ser5 phosphorylation levels (compare lane 2 versus lane 3). Therefore, the observed effect of FKBP5 knockdown on global increases in Pol II Ser2 and Ser5 phosphorylation is independent of the functions of HSP90. Furthermore, in support of this HSP90-independent effect of FKBP5, we also failed to observe any increase in CDK9 interaction with HSP90 upon FKBP5 knockdown compared to that in control scramble cells (Fig. 4G, compare lane 3 versus lane 4).

FKBP5 and cyclin T1 interact with CDK9 through the same domain.

Our earlier results have shown that the inhibitory effect of FKBP5 on P-TEFb-mediated phosphorylation activity depends at least in part on its interaction with CDK9 both in vitro and in vivo within mammalian cells. Towards a deeper mechanistic understanding of this regulation, we initially addressed the specific region within CDK9 that would interact with FKBP5. For providing direct evidence, we initially addressed the CDK9 domain that would interact with FKBP5 through their expression in heterologous Sf9 cells. We generated several plasmids expressing CDK9 fragments deleted at both N and C termini and expressed them through baculovirus-mediated infection of Sf9 cells (see Fig. 5A and B for fragments used). As shown in Fig. 5B, coinfection of these fragments along with full-length FKBP5 showed that while the full-length CDK9 strongly interacts, a deletion of 31 amino acids from the N-terminal end almost abolished CDK9 interaction with FKBP5 (compare lane 5 versus lane 6). Since the N-terminal half of CDK9 was reported previously to interact with cyclin T1 (44), we also addressed whether this amino acid region (1 to 31) of CDK9 would also be required for its interaction with cyclin T1. As shown in Fig. 5C, immunoprecipitation analyses using extract of Sf9 cells expressing the indicated CDK9 fragments and cyclin T1 clearly showed dependence on the same region of CDK9 for its interaction with cyclin T1 within this heterologous system (compare lane 5 versus lane 6). Furthermore, we also addressed whether the same domain of CDK9 would be required for its interaction with both partners within mammalian cells. We transfected 293T cells expressing the indicated CDK9 fragments and tested their interaction with endogenous cyclin T1 and FKBP5. As shown in Fig. 5D and consistent with our observations in heterologous Sf9 cells, we also observed a strong dependence on the N-terminal 31 amino acids for CDK9 interaction with both cyclin T1 and FKBP5 (compare lane 10 versus lane 11). Thus, based on all of this evidence, we conclude that the N-terminal 31 amino acids of CDK9 are required for its interactions with both cyclin T1 and FKBP5. These results also further indicate a possibility of existence of competition between FKBP5 and cyclin T1 for their binding at the N terminus of CDK9, thus providing a mechanistic explanation of the negative role of FKBP5 on P-TEFb-mediated phosphorylation both in vitro and in vivo within mammalian cells.

FIG 5.

FKBP5 and cyclin T1 interact in the same N-terminal region of CDK9. (A) Cartoon diagram representing various functional domains within CDK9 that are important for its functions. NLD, nuclear localization domain. Amino acids 26 to 31, ATP-binding site; amino acids 60 to 66, PITALRE; amino acids 1 to 82, p53-binding site. (B) Domain analysis of CDK9 (using various domains as indicated) for its interaction with FKBP5 through their baculovirus-mediated expression in Sf9 cells. Sf9 cells were coinfected with the baculoviruses as indicated, and cell lysates were subjected to immunoprecipitation and subsequent Western blotting for identifying their interaction within this heterologous system. (C) Domain analysis of CDK9 (using various domains as indicated) for its interaction with cyclin T1 through their baculovirus-mediated expression in Sf9 cells. Sf9 cells were coinfected with the baculoviruses as indicated, and cell lysates were subjected to immunoprecipitation and subsequent Western blotting for identifying their interaction within this heterologous system. (D) Domain analysis of CDK9 (using various domains as indicated) for its interaction with endogenous FKBP5 and cyclin T1 within mammalian 293T cells. 293T cells were transfected with various CDK9 plasmids as indicated, and cell lysates were subjected to immunoprecipitation and subsequent Western blotting using factor-specific antibodies for identifying their interaction within 293T cells.

FKBP5 and cyclin T1 compete with each other for their binding to CDK9.

Since our earlier analyses indicated the presence of competition between FKBP5 and cyclin T1 for their binding to CDK9 owing to their use of the same binding sites (Fig. 5), we subsequently addressed this competitive mode of binding between these factors in details through several experiments. In the initial analysis, we cotransfected 293T cells with plasmids expressing CDK9 and increasing concentrations of FKBP5 to test whether its increased expression and subsequent binding would reduce concomitant CDK9 binding to cyclin T1 within mammalian cells. As shown in Fig. 6A, ectopically expressed CDK9 efficiently bound endogenous cyclin T1 (lane 3). Interestingly, and consistent with our hypothesis of a competition model, increased expression, and thus increased binding, of FKBP5 concomitantly reduced CDK9 interaction with endogenous cyclin T1 (Fig. 6A, compare lane 3 versus lanes 4 and 5). Furthermore, this reduction in interaction requires FKBP5 binding to CDK9, since an FKBP5 fragment (1 to 157 amino acids) that failed to show an interaction with CDK9 (Fig. 2D and F), also failed to show an effect on CDK9 binding with endogenous cyclin T1 (Fig. 6B, compare cyclin T1 binding between lanes 3 and4 versus 5 and 6). In a reverse experiment, increased expression of ectopic cyclin T1 and its increased binding to CDK9 resulted in concomitant reduced binding of FKBP5 as well (Fig. 6C, compare lane 3 versus lanes 4 and 5) within mammalian cells.

FIG 6.

FKBP5 and cyclin T1 compete with each other for their binding to CDK9 both in vitro and in vivo within mammalian cells. (A) Immunoblotting analysis showing the effect of increasing expression of FKBP5 on the association of ectopically expressed CDK9 and endogenous cyclin T1. 293T cells were cotransfected with the indicated plasmids, and cell lysates were subjected to immunoprecipitation using epitope tag-specific agarose beads and to subsequent Western blotting using factor-specific antibodies for identifying their interactions. (B) Immunoblotting analysis showing effect of expression of full-length (FL) and CDK9 interaction-defective (1 to 157 amino acids) FKBP5 proteins in 293T cells on association of ectopically expressed CDK9 with endogenous cyclin T1. 293T cells were cotransfected with the indicated plasmids, and cell lysates were subjected to immunoprecipitation using epitope tag-specific agarose beads and to subsequent Western blotting using factor-specific antibodies for identifying their interactions. (C) Immunoblotting analysis showing effect of increasing expression of cyclin T1 on association of ectopically expressed CDK9 with endogenous FKBP5. 293T cells were cotransfected with the indicated plasmids, and cell lysates were subjected to immunoprecipitation using epitope tag-specific agarose beads and to subsequent Western blotting using factor-specific antibodies for identifying their interaction. (D) SDS-PAGE Coomassie staining showing purification of GST-cyclin T1 used in panels E to G. A filled red dot indicates the target protein band, whereas others represent either degradation or nonspecific proteins. (E) Immunoblotting analysis showing direct interaction of CDK9 with cyclin T1. Purified recombinant CDK9 and cyclin T1 proteins were used and added as indicated in the in vitro interaction assays. (F) Immunoblotting analysis showing effect of increasing addition of FKBP5 on association of CDK9 with cyclin T1 in vitro. Purified recombinant CDK9, cyclin T1, and FKBP5 proteins were used and added as indicated in the in vitro interaction assays for identifying the effect of increasing addition of FKBP5 on CDK9 association with cyclin T1 in vitro. For each reaction mixture, 60 μM purified FLAG-CDK9, 50 μM GST-cyclin T1 in each lane as appropriate, and 100 μM, 100 μM, and 200 μM (in lanes 3, 4, and 5, respectively) purified FKBP5 were added. (G) Immunoblotting analysis showing the effect of increasing addition of cyclin T1 on association of CDK9 with FKBP5 in vitro. Purified recombinant CDK9, cyclin T1, and FKBP5 proteins were used and added as indicated in the in vitro interaction assays for identifying the effect of increasing addition of cyclin T1 on CDK9 association with FKBP5 in vitro. For each reaction mixture, 60 μM purified FLAG-CDK9, 50 μM His-FKBP5 in each lane as appropriate, and 50 μM, 50 μM, and 100 μM (in lanes 3, 4, and 5, respectively) purified GST-cyclin T1 were added. (H) Immunoblotting analysis showing effect of FKBP5 knockdown on association of endogenous CDK9 with cyclin T1. Endogenous CDK9 was immunoprecipitated using specific antibody, and its association with endogenous cyclin T1 was assessed by Western blot analysis. (I) (Left) Experimental strategy that was employed for our assay as shown in the right subpanel. (Right) In vitro dissociation assay demonstrating the effect of addition of purified recombinant FKBP5 in dissociating cyclin T1 from preformed P-TEFb complex obtained through recombinant expression of FLAG-CDK9 and GST-cyclin T1. It can be observed that purified FKBP5 efficiently dissociates cyclin T1 from CDK9, and thus the level of cyclin T1 is increased in the supernatant and is decreased with bead-bound CDK9.

For providing direct evidence of this competitive binding, we performed in vitro interaction analyses with purified FKBP5, CDK9, and cyclin T1 proteins (Fig. S2A and B and Fig. 6D). The cyclin T1 protein was purified through its expression in a bacterial system. Our initial analysis with purified CDK9 and cyclin T1 clearly showed their direct interaction in vitro (Fig. 6E, lane 4). Interestingly, addition of purified FKBP5 to this interaction assay clearly showed reduced cyclin T1 binding to CDK9 upon concomitant increased binding of FKBP5 (Fig. 6F, compare cyclin T1 binding to CDK9 in lane 2 versus lane 4 and 5). In a reciprocal experiment, increased addition of cyclin T1 and its increased binding also reduced concomitant FKBP5 binding to CDK9 (Fig. 6G, compare FKBP5 binding to CDK9 in lane 2 versus lanes 4 to 5).

To provide further evidence of this model of competitive binding in an endogenous context, we used FKBP5 knockdown cells (Fig. 4A). Immunoprecipitation of endogenous CDK9 clearly showed enhanced CDK9 binding with endogenous cyclin T1 upon FKBP5 knockdown compared to that upon control scramble knockdown (Fig. 6H, compare lane 2 versus lane 3). For providing further mechanistic evidence of these functional regulations, we used purified preformed P-TEFb complex (FLAG-CDK9 plus GST-cyclin T1) in our in vitro dissociation assay (see Materials and Methods). As shown in Fig. 6I, addition of purified recombinant FKBP5 efficiently dissociated cyclin T1 from the preformed P-TEFb complex such that its overall amount in the supernatant increased with increasing concentration of added FKBP5 (compare lane 5 with lanes 6 to 8). Consistent with this, cyclin T1 bound to FLAG-CDK9 also concomitantly decreased with increasing addition of FKBP5 (Fig. 6I, lane 1 versus lanes 2 to 4). Furthermore, FKBP5 could also efficiently dissociate cyclin T1 from immunoprecipitated P-TEFb complex from mammalian cells upon its binding with CDK9 (Fig. S5A). These data provide evidence of FKBP5 directly modulating cyclin T1 association with CDK9 that can have implications in overall P-TEFb-mediated functional regulations both in vitro (Fig. 3) and in vivo within mammalian cells (Fig. 4). Based on all of this evidence, we conclude that human FKBP5 and cyclin T1 compete with each other for their binding to CDK9 (Fig. S5B).

FKBP5 regulates expression of diverse sets of genes.

Based on our mechanistic understanding of the role of FKBP5 in negative regulation of P-TEFb functions, we were further interested in understanding the implication of this regulation on expression of global target genes. We performed transcriptome sequencing (RNA-Seq) analysis for identifying genes that would be affected upon FKBP5 knockdown (GEO accession no. GSE169461). For the purpose of control scramble cells, we used data sets that were reported in our earlier study (37) (SRA accession no. PRJNA512165), since all of these experiments were performed at the same time. Transcriptome analysis revealed a rather small set of genes (a total of 219) showing significant (P < 0.05) upregulation of expression (log₂ fold change [log₂FC] > 1.5) and a similar number of genes (a total of 145) being downregulated upon FKBP5 knockdown. A heatmap representing expression of some of these genes is shown in Fig. 7A. Interestingly, subsequent gene ontology analyses showed a strong association of some of these upregulated genes with multiple cellular processes, such as Pol II-mediated transcriptional regulation, as well as in cellular differentiation and proliferation-related events (Fig. 7B). In contrast, the downregulated genes showed a strong association with regulation of multiple metabolic processes. The overall effect of FKBP5 knockdown on global target gene expression further indicates a role for FKBP5 in regulation of Pol II-mediated transcriptional processes for expression of genes that are important for regulation of cellular proliferation and differentiation.

FIG 7.

FKBP5 knockdown enhances expression of proliferation-related genes. (A) Heatmap showing the effect of FKBP5 knockdown on expression of selected target genes compared to that in control scramble cells. Log₂FC represents fold change on a log₂ scale. (B) Gene ontology (GO) analysis predicting association of upregulated and downregulated genes upon FKBP5 knockdown with various biological processes. (C) qRT-PCR analysis confirming upregulation of mRNA expression of target genes that are involved in cellular proliferation in 293T cells upon FKBP5 knockdown. Data represent n = 2 biological replicates and 3 PCR replicates of each biological replicate. (D) qRT-PCR analysis showing effect of reexpression of full-length and CDK9 interaction-defective FKBP5 proteins in FKBP5 knockdown cells on mRNA expression of indicated target genes. Data represent n = 2 biological replicates and 3 PCR replicates of each biological replicate. (E) Chromatin immunoprecipitation (ChIP) analysis showing enhanced recruitment of the P-TEFb complex (using CDK9 as the candidate) upon FKBP5 knockdown resulting in enhanced level of Pol II CTD Ser2 and Ser5 phosphorylation and reduced levels of Pol II in the promoter-proximal region of CCND1 and C-MYC target genes. Data represent n = 2 biological replicates and 3 PCR replicates of each biological replicate. (F) Quantitation of relative enrichment of Pol II CTD Ser2 and Ser5 phosphorylation normalized with the amount of total Pol II being present on the promoter-proximal regions of target CCND1 and C-MYC genes. Data represent n = 2 biological replicates and 3 PCR replicates of each biological replicate. (G) ChIP analysis showing enhanced presence of Pol II at the indicated coding regions of target CCND1 and C-MYC genes. Data represent n = 2 biological replicates and a minimum of 3 PCR replicates of each biological replicate. In all of our assays, statistical analyses were performed using a one-tailed Student’s t test for obtaining statistical significance of our data. *, P ≤ 0.05; **, P ≤ 0.01; ***; P ≤ 0.001; ns, not significant.

Knockdown of FKBP5 enhances expression of proliferation-related genes.

Based on our RNA-Seq analysis and an indication of involvement of a subset of these upregulated genes in controlling cellular proliferation, we were interested in identifying molecular mechanism of the upregulation of expression upon FKBP5 knockdown. Subsequent quantitative real-time PCR (qRT-PCR) analyses confirmed significant upregulation of genes that are widely involved in regulation of cellular proliferation upon FKBP5 knockdown (Fig. 7C). Expression of two of these critical genes, CCND1 and C-MYC, encoding cyclin D1 and c-Myc proteins, respectively, have been widely studied for their role as master regulators of expression of global target genes. Along with these two critical regulators, we also observed upregulation of other key regulators of cellular proliferation, including CDKN1A, CDKN1B, and MDM2, among others. This upregulation is specific, since in the same assay, we failed to observe any effect of FKBP5 knockdown on expression of nontarget genes such as RGS2, MAP2, and LDLR. Similar effects are also observed upon FKBP5 knockdown on target gene expression in a prostate cancer cell line (PC3) (Fig. S6A), which thus rules out an effect of FKBP5 knockdown on increasing expression of proliferation-related genes in a cell type-specific manner. The overall effect of FKBP5 knockdown in increasing the expression of target genes is specific, since reexpression of full-length FKBP5 reversed the effect, as shown for two target genes (CCND1 and E2F2), within 293T cells, whereas the CDK9 interaction-deficient FKBP5 mutant (1 to 157 amino acids) failed to do so (Fig. 7D). To further corroborate the overall effect, we tested the expression level of these target genes in CDK9 knockdown cells (Fig. S6B). As shown in Fig. S6C, the majority of these target genes are also downregulated upon CDK9 knockdown, indeed suggesting the overall effect of FKBP5 knockdown on target gene expression through modulation of P-TEFb activity. Furthermore, overexpression of CDK9 that increased Pol II CTD phosphorylation (Fig. S6D) also caused significant increases in expression of all these target genes (Fig. S6E) and provided additional evidence of functional effect of FKBP5 in modulating P-TEFb-dependent target gene expression within mammalian cells. Also, consistent with a role of P-TEFb in stimulating transcriptional activation of HIV long terminal repeat (LTR) promoter-driven genes, we also observed an increase in the overall expression of HIV LTR promoter-driven luciferase reporter gene upon FKBP5 knockdown compared to that in control scramble cells (Fig. S6F).

FKBP5 knockdown enhances recruitment of P-TEFb complex at the target CCND1 and C-MYC genes.

Once we established CCND1 and C-MYC genes as key targets whose expressions are regulated by FKBP5, we were interested in further mechanistic understanding of this regulation. Both of these genes have been shown to be key targets for P-TEFb complex-containing SEC for transcriptional activation (37, 45). Since FKBP5 negatively regulates P-TEFb level and its functions within the cellular system, we wondered whether enhanced recruitment of the P-TEFb complex upon FKBP5 knockdown would correlate with enhanced expression of these genes. As shown in Fig. 7E, our chromatin immunoprecipitation (ChIP) analysis showed a significant increase in CDK9 (representing P-TEFb complex) recruitment at the promoter-proximal region of these target genes. An increase in P-TEFb complex recruitment also results in increased presence of Pol II CTD Ser2 and Ser5 phosphorylation (Fig. 7E), whereas similar analyses with the control RGS2 gene failed to show any effect. Consistent with a role for P-TEFb-mediated phosphorylation in releasing paused Pol II from the promoter-proximal region, we also observed a significant decrease in the overall level of total Pol II at the promoter-proximal region of these genes. Enhanced P-TEFb recruitment in the FKBP5 knockdown cells compared to that in the control scramble cells (Fig. 7F). Consistent with decreases in Pol II levels at the transcriptional start site (TSS) region, we also observed an increased presence of Pol II in the coding region of both of these genes (Fig. 7G). Pausing index analysis clearly showed significant reduction in overall pausing of Pol II at both of these target genes upon FKBP5 knockdown (Fig. S6G). Therefore, based on all of this evidence, we conclude that human FKBP5 negatively regulates the overall level of the P-TEFb complex within mammalian cells for regulating expression of key proliferation-related target genes such as CCND1 and C-MYC. These results also point out to an antiproliferative role of FKBP5 in regulating cell growth.

Knockdown of FKBP5 enhances cellular proliferation.

Since our study has shown a negative role of FKBP5 in P-TEFb-mediated expression of proliferation-related genes, we anticipated that reduced expression of FKBP5 might enhance overall cellular proliferation. Indeed, consistent with this hypothesis, knockdown of FKBP5 significantly enhanced proliferation of 293T cells compared to control scramble knockdown (Fig. S6H). Consistent with a similar effect of FKBP5 knockdown on expression of proliferation-related genes in the PC3 prostate cancer cell line (Fig. S6A), we also observed enhanced cellular proliferation of the PC3 cell line upon FKBP5 knockdown (Fig. S6I), thus ruling out an overall effect of FKBP5 in controlling proliferation in a cell-type-specific manner.

Strong association of downregulation of expression of FKBP5 and upregulation of expression of proliferation-related genes in several human cancers.

Since our results clearly indicated an antiproliferative role of FKBP5 in controlling cell growth, we assumed that similar functional significance would also be important for cancer pathogenesis. It is of note that a few earlier studies have indicated a negative correlation between FKBP5 expression and severity of cancers (8, 19). To address FKBP5 functions and target gene expression in cancer pathogenesis, we analyzed mRNA expression of FKBP5 and its correlation with expression of proliferation-related genes in patient cohort samples of multiple cancer types available in The Cancer Genome Atlas (TCGA) data sets. We were intrigued by our observation of significant correlation between downregulation of FKBP5 expression and the increased expression of proliferation-related genes such as CCND1, and C-MYC (Fig. S7A to E). Thus, based on mechanistic understanding of the functional role of FKBP5 as deciphered by our study, downregulation of FKBP5 expression could play an essential role in overall cancer pathogenesis through enhanced expression of key cancer-causing protooncogenes such as CCND1 and C-MYC. Furthermore, it is interesting to note that several forms of kidney cancer show strong correlations between downregulation of FKBP5 expressions and upregulation of CCND1 and C-MYC genes, which further substantiates our observation in 293T cells. Our observation that FKBP5 downregulation enhances cell proliferation through enhancement of expression of proliferation-related genes further supports the overall role of FKBP5 as a tumor suppressor in these types of cancers.

DISCUSSION

One of the key findings of our study is to decipher a novel function of human FKBP5 in regulation of Pol II-mediated transcription through regulation of P-TEFb levels within mammalian cells. Owing to having the same binding sites on CDK9, FKBP5 competes with cyclin T1 for their interaction and thus regulates the level of P-TEFb complex formation through regulation of cyclin T1 association with CDK9 for transcriptional activation. Consistent with this model, overexpression of FKBP5 reduces CDK9 association with cyclin T1 and transcriptional activation, whereas knockdown of FKBP5 expression increases CDK9-cyclin T1 association and causes transcriptional activation. A working model, as deciphered by our studies, is presented in Fig. S8 in the supplemental material.

Potential implication of role of FKBP5 in regulating pausing events at the promoter-proximal region.

An earlier study reported a role for the HSP90 chaperone protein in regulating pausing events at the promoter-proximal region of global target genes (46). This study indicated a role for HSP90 interaction with the NELF complex in overall regulation of pausing without providing deeper mechanistic insights. Since FKBP5 plays a key role as a cochaperone of the HSP90 chaperone complex, it is highly likely that, through the mechanisms as deciphered in our study, FKBP5 could also play additional roles within the HSP90 chaperone complex for attaining overall pausing event through negative regulation of functions of the P-TEFb complex which is important for releasing the paused Pol II from the promoter-proximal region. Consistent with this hypothesis, our ChIP analysis of target CCND1 and C-MYC genes showed a significant decrease in overall Pol II levels upon knockdown of FKBP5 at the promoter-proximal region (Fig. 7E). Further studies are needed for detailed understanding of this mechanism of global pausing event involving FKBP5 within mammalian cells.

FKBP5-mediated transcriptional downregulation and cancer pathogenesis.

It is quite interesting to note that in our study, significant number of cancers show a strong correlation between downregulation of FKBP5 and increased expression of key proliferation-related genes (see Fig. S4 in the supplemental material). Mechanistic understanding obtained from our study provides a paradigm for this overall observation, in which enhanced activation of P-TEFb functions upon reduced FKBP5 expression causes upregulation of expression of protooncogenes such as C-MYC, CCND1, etc. Overactivation of AKT functions is also implicated in breast cancer pathogenesis (8). A strong correlation between downregulation of FKBP5 expression and enhanced expression key proliferation-related genes in multiple cancers further provides a functional role of FKBP5 as a tumor suppressor in multiple types of cancers. The combined role of FKBP5-mediated AKT and P-TEFb functions in overall cancer pathogenesis is a subject for future studies to come.

Transcriptional downregulation by FKBP5 and type 2 diabetes pathogenesis.

The human FKBP5 has also been implicated in glucose metabolism and associated type 2 diabetes pathogenesis. FKBP5 knockout mice show improved glucose tolerance associated with leaner body weight when fed with normal chow diet (47, 48). Interestingly, a recent study has shown a role for the AKT signaling pathway, upon insulin stimulation, in phosphorylation of target AS160, which helps in translocation of the GLUT4 glucose uptake protein to the plasma membrane and thus facilitates overall glucose uptake and homeostasis (15). Interestingly, the same study also observed an increase in GLUT4 expression upon ablation of FKBP5 functions by its knockdown or pharmacological antagonist. The results, as deciphered in our study, provide a new mechanistic insight into this overall regulation. It is quite possible that upon FKBP5 knockdown, the overall increase in P-TEFb functions results in increased transcription of GLUT4, leading to its increased expression. Whether pharmacological inhibition would also result in a similar increase in P-TEFb functions remains to be studied. Nevertheless, the important mechanistic insight on functional regulation of FKBP5 would provide a basis for pathogenesis of multiple diseases, including that of type 2 diabetes.

Transcriptional downregulation by FKBP5 and adipogenesis.

Few studies have implicated a role for FKBP5 in regulation of adipogenesis. The maximum amount of FKBP5 expression has been reported in adipocytes, as well as in skeletal muscles and lymphocytes, compared to that in other tissues (49). Therefore, not surprisingly, FKBP5 functions have been correlated with adipocyte differentiation from white adipose tissue (WAT) to brown adipose tissue (BAT) (18). FKBP5 knockdown has been shown to increase in expression of the key regulators of adipogenesis UCP1 and PRDM16 (18). However, the underlying mechanisms of this increase are completely unknown. The overall role of FKBP5 in transcriptional regulation through negative regulation of functions of the P-TEFb complex certainly provides a mechanistic explanation of this overall regulation. Functions of the P-TEFb complex have been shown to be involved in regulation of expression of multiple key genes during WAT to BAT differentiation. In this context, whether FKBP5 downregulation and its implication in an increase in P-TEFb-mediated transcriptional upregulation of key UCP1 and PRDM16 genes, as well as others, plays an important role in overall adipogenesis would be a subject for detailed studies.

Functions of FKBP5 and neuronal pathogenesis.

Consistent with its high expression in the hippocampus and amygdala (50), the two key regions regulating stress-related functions, FKBP5 has been implicated in controlling the pathogenesis of several neuronal disorders, including PTSD, by a significant number of studies (4–6). The majority of these studies have implicated a role for FKBP5 in regulating the glucocorticoid receptor (GR) signaling pathway, in which FKBP5 competes with glucocorticoids for their binding to the GR and subsequent downstream signal transduction for transcriptional activation of target genes. However, no other studies have provided direct mechanistic insights into the functional role of FKBP5 in transcription. Thus, our study has the potential to decipher important mechanistic insights into the role of FKBP5 and overall neuronal pathogenesis that studies have failed to explore until now. Further future studies would be directed toward addressing some of these key questions.

Overall, the mechanistic insights obtained from our study have the potential to increase understanding of underlying mechanisms of pathogenesis of several FKBP5-associated diseases, including cancer and metabolic disorders, as well as several neuronal disorders that studies until now have not been able to address.

MATERIALS AND METHODS

Cell culture.

For this study, mammalian cells were cultured in Dulbecco’s modified Eagle medium (DMEM) with high glucose (Gibco, USA), supplemented with 10% fetal bovine serum (FBS; Gibco) and 100 U/mL penicillin-streptomycin (Gibco). Mammalian cells were maintained in a 5% CO₂ incubator at 37°C while maintaining proper humid conditions. All insect cell (Sf9) culturing in this study was done in Grace’s insect medium (HiMedia, India) supplemented with 10% FBS (Gibco) and 7 µg/mL gentamycin. Sf9 cells were maintained in a bio-oxygen demand (BOD) incubator at 26°C.

Transfection of mammalian and Sf9 cells.

Transfection of mammalian cells were done using FuGene transfection reagent and following the manufacturer’s protocol. Unless otherwise mentioned, cells were harvested 48 h after transfection for subsequent assays. To generate baculoviruses for expression of recombinant proteins, Sf9 cells were transfected using Cellfectin II reagent (Invitrogen) per the manufacturer’s protocol. Virus particles (supernatant) were collected 72 h posttransfection.

Plasmids, primers, and antibodies.

All the plasmids used for this study, primers used for RNA analyses, ChIP analyses, construction of shRNAs, and antibodies used for various experiments are mentioned in Tables 1 to 5.

TABLE 1.

List of plasmids used in this study

Plasmid name	Description
M219	HA-tagged NELF-A gene cloned into FLAG-pFASTBAC vector
M220	Myc-tagged NELF-B gene cloned into FLAG-pFASTBAC vector
M221	VSVG-tagged NELF-C gene cloned into FLAG-pFASTBAC vector
M222	FLAG-tagged NELF-E gene cloned into FLAG-pFASTBAC vector
M250	CDK9 cloned into FLAG-HA pCDNA5-FRT-TO vector
M384	FKBP5 gene cloned in pCMVSPORT6 vector
M250	CDK9 cloned into FLAG-HA pCDNA5-FRT-TO vector
M452	FKBP5 cloned into FLAG-HA pCDNA5-FRT-TO vector
M493	FKBP5 cloned into 6-His pET-11d vector
M494	FKBP5 cloned into pFASTBAC vector plasmid containing FLAG tag
M495	FKBP5 cloned into pFASTBAC vector plasmid containing Non tag
M502	RNA Pol II CTD cloned into pGEX vector
M503	FKBP5 cloned into FLAG-tagged pcDNA5-FRT-TO vector
M504	FKBP5 cloned into HA-tagged pcDNA5-FRT-TO vector
M619	CDK9 cloned into FLAG-tagged pcDNA5-FRT-TO vector
M620	CDK9 cloned into HA-tagged pcDNA5-FRT-TO vector
M661	FKBP5 (1–384) fragment cloned into FLAG pCDNA5-FRT-TO vector
M662	FKBP5 (1–351) fragment cloned into FLAG pCDNA5-FRT-TO vector
M663	FKBP5 (1–317) fragment cloned into FLAG pCDNA5-FRT-TO vector
M664	FKBP5 (1–268) fragment cloned into FLAG pCDNA5-FRT-TO vector
M665	FKBP5 (1–157) fragment cloned into FLAG pCDNA5-FRT-TO vector
M666	FKBP5 (130–452) fragment cloned into FLAG pCDNA5-FRT-TO vector
M667	FKBP5 (243–452) fragment cloned into FLAG pCDNA5-FRT-TO vector
M411	CDK9 cloned into FLAG-pFASTBAC vector
M251	CDK9 cloned into Non tag pFASTBAC vector
M296	psPAX2 lentivirus packaging plasmid from Addgene
M297	pMD2.G lentivirus envelope plasmid from Addgene
M298	Lentiviral pLKO.1 vector containing scramble sequence
M562	Cyclin T1 cloned into pGEX vector
M669	FKBP5 gene cloned into Gal-DBD pM vector
M781	FKBP5 (1–384) fragment cloned into FLAG-pFASTBAC vector
M782	FKBP5 (1–351) fragment cloned into FLAG-pFASTBAC vector
M783	FKBP5 (1–268) fragment cloned into FLAG-pFASTBAC vector
M784	FKBP5 (1–157) fragment cloned into FLAG-pFASTBAC vector
M785	FKBP5 (130–452) fragment cloned into FLAG-pFASTBAC vector
M786	FKBP5 (243–452) fragment cloned into FLAG-pFASTBAC vector
M808	Human CDK7 gene cloned into FLAG-pFASTBAC vector
M809	Human CDK8 gene cloned into FLAG-pFASTBAC vector
M916	CDK9 (1–276) fragment cloned into FLAG-pFASTBAC vector
M917	CDK9 (32–276) fragment cloned into FLAG-pFASTBAC vector
M918	CDK9 (67–276) fragment cloned into FLAG-pFASTBAC vector
M919	CDK9 (100–276) fragment cloned into FLAG-pFASTBAC vector
M920	CDK9 (150–276) fragment cloned into FLAG-pFASTBAC vector
M928	FKBP5 (1–384) fragment cloned into 6-His pET-11d vector
M929	FKBP5 (1–351) fragment cloned into 6-His pET-11d vector
M930	FKBP5 (1–317) fragment cloned into 6-His pET-11d vector
M931	FKBP5 (1–268) fragment cloned into 6-His pET-11d vector
M932	FKBP5 (1–157) fragment cloned into 6-His pET-11d vector
M933	FKBP5 (130–452) fragment cloned into 6-His pET-11d vector
M934	FKBP5 (243–452) fragment cloned into 6-His pET-11d vector
M935	CDK9 (1–276) fragment cloned into FLAG-tagged pcDNA5-FRT-TO vector
M936	CDK9 (32–276) fragment cloned into FLAG-tagged pcDNA5-FRT-TO vector
M937	CDK9 (67–276) fragment cloned into FLAG-tagged pcDNA5-FRT-TO vector
M938	CDK9 (100–276) fragment cloned into FLAG-tagged pcDNA5-FRT-TO vector
M939	CDK9 (150–276) fragment cloned into FLAG-tagged pcDNA5-FRT-TO vector
M341	pRL-TK-internal Renilla luciferase control plasmid
M1008	pLTR-Luc (with HIV LTR promoter) in pGL3 basic vector
S110	shRNA construct cloned into pLKO.1 for human FKBP5 no. 1
S111	shRNA construct cloned into pLKO.1 for human FKBP5 no. 2
S142	shRNA construct cloned into pLKO.1 for human CDK9 no. 1

TABLE 2.

List of primers used for RNA analysis

Gene name	5′ forward primer	3′ reverse primer
FKBP5	TCCTTGCTGCCTTTCTG	CTTTCTCAAAGTCACCCTTG
CDK9	GCATCATGGCAGAGATGTG	GTTGTCCACGTTTGGCC
CCND1	CAAACACGCGCAGACCTTC	GATCACTCTGGAGAGGAAGCG
C-MYC	GCTTGTACCTGCAGGATC	GACTCCGTCGAGGAGAG
CDKN1A	GGACAGCAGAGGAAGACCATG	CTGTCATGCTGGTCTGCC
CDKN1B	CGACGATTCTTCTACTCAA	TTACGTTTGACGTCTTCTG
CDKN1C	GCTGCACTCGGGGATTTC	GGACATCGCCCGACGACT
CDC25B	GGCACATCAAGACTGCGG	GGTAGTCGTTGACAGCACG
ATF2	GTACCAGGCCCATTTCCTCTTC	GAACGAGTGGGACTGCAGCTG
E2F2	GGAGCCGGACAGTCCTTC	GCTGTCAGTAGCCTCCAAG
MAP2	AATAGACCTAAGCCATGTGACATCC	AGAACCAACTTTAGCTTGGGCC
TAF4	GACGACAGATATGAGCAGG	GTTGCTGCATCTCCTTTG
PTPRS	AACACAGAAGTGCCCGCAC	GTGACGTGTGGGCCTTGGAG
CDK6	GGAGTGTTGGCTGCATATTTG	CGATATCTGTTACAAACTTC
TM2D2	GGACACTACTTCATAACCAC	AATAAGGTCAACAAACCACC
MDM2	TTGGATCAGGATTCAGTTTC	GAGAGTTCTTGTCCTTCTTC
TEF	GCTCTTCCACAGCATCC	GCACACTGGAGAGCAC
LDLR	CTGGAAATTGCGCTGGAC	GTCTTGGCACTGGAACTCGT
GAPDH	CATCACCATCTTCCAGGAG	GTTCACACCCATGACGAAC
18S rRNA	GTAACCCGTTGAACCCCATT	CCATCCAATCGGTAGTAGCG
RGS2	AAGATTGGAAGACCCGTTTGAG	GCAAGACCATATTTGCTGGCT
STIP1	GAGAAAATCCTGAAGGAGCAAG	ATGCTTCATGGCCTGGGGATA
GLUT1	GAAGTGTCACCCACAGCCCTTC	GGGCCACAGGTCCTTGTTGC
CRABP2	CCCTACACCAACAAAGAGG	CCCTCAAGTCCCCTTTAG

TABLE 3.

List of primers used in ChIP analyses

Gene name	5′ forward primer	3′ reverse primer
CCND1 (TSS)	CGGGCTTTGATCTTTGCTTA	CTGCTGCTCGCTGCTACT
C-MYC (TSS)	TCCTCTCTCGCTAATCTCCGC	GGGTCCTCAGCCGTCCAGAC
RGS2 (TSS)	CGAGGCCTCATAAATGCTG	CTCTTGTCCATGGGTCTG
CCND1 (∼2 kb downstream)	CTGTTTTGATCTGGGATTGCGTGT	CACCTCCGGCTCAGAGCACTG
C-MYC (∼3 kb downstream)	GACTCTGGTAAGCGAAGC	TCCAGATCTGCTATCTCTCC

TABLE 4.

Oligonucleotide sequences for generation of shRNA constructs

Gene	Sequence
	Upper	Lower
FKBP5 sh#1	CCGGACCTAATGCTGAGCTTATATACTCGAGTATATAAGCTCAGCATTAGGTTTTTTG	AATTCAAAAAACCTAATGCTGAGCTTATATACTCGAGTATATAAGCTCAGCATTAGGT
FKBP5 sh#2	CCGGCGAAGGAGCAACAGTAGAAATCTCGAGATTTCTACTGTTGCTCCTTCGTTTTTG	AATTCAAAAACGAAGGAGCAACAGTAGAAATCTCGAGATTTCTACTGTTGCTCCTTCG
CDK9 sh#1	CCGGCCGCTGCAAGGGTAGTATATACTCGAGTATATACTACCCTTGCAGCGGTTTTTG	AATTCAAAAACCGCTGCAAGGGTAGTATATACTCGAGTATATACTACCCTTGCAGCGG

TABLE 5.

List of antibodies used in this study

Antibody	Source(s)
FKBP5	Bethyl Laboratories, Cell Signaling Technology
CDK9	Santa Cruz Biotechnology
ELL	Bethyl Laboratories, Cell Signaling Technology
CCNT1	Santa Cruz Biotechnology
AFF1	Abcam
NELF-A	Bethyl Laboratories
NELF-E	Bethyl Laboratories
AF9	Bethyl Laboratories
β-Actin	Santa Cruz Biotechnology
Phospho-Rpb1 CTD (Ser2)	Cell Signaling Technology
Phospho-Rpb1 CTD (Ser5)	Cell Signaling Technology
Rpb1	Cell Signaling Technology
FLAG epitope	Sigma
HA epitope	Santa Cruz Biotechnology
His epitope	Santa Cruz Biotechnology
GST epitope	Santa Cruz Biotechnology
Normal rabbit IgG	Cell Signaling Technology
Normal mouse IgG	Cell Signaling Technology
HRP-conjugated secondary antibody (rabbit)	Bio-Rad
HRP-conjugated secondary antibody (mouse)	Cell Signaling Technology
HSP90β	Bio-Rad
Histone H3	Cell Signaling Technology

Construction of plasmids used in this study.

We obtained pCMVSPORT6 vector, with the FKBP5 gene cloned into it, from Open Biosystems. The construct was sequenced in its entirety before being used for downstream cloning experiments. Full-length as well as derivatives of different FKBP5 constructs were cloned in multiple expression vectors, including pcDNA5/FRT/TO, pFASTBac, pET-28a(+), and pGEX. Similarly, CDK9, cyclin T1, CDK7, CDK8, Pol II CTD, and other required factors and their fragments were PCR amplified and cloned in different vectors for their expression with epitope tags and in mammalian, Sf9, and bacterial cells. The details of cloning information for any given construct are available through the corresponding author of this study.

Immunoprecipitation analysis.

To test the interaction between FKBP5 and CDK9 in an endogenous context, nuclear extracts were prepared from a confluent 10-cm dish of 293T cells. Subsequently, this extract was precleared using protein A-agarose beads for 2 h at 4°C. The precleared extract was subsequently used for overnight immunoprecipitation with 2 µg of indicated antibody. Antibody-bound proteins were pulled down using protein A-agarose beads by incubating for 4 h. After rigorous washing, bound proteins were eluted by boiling in 1× SDS loading dye for 10 min at 95°C. The sample was analyzed by Western blotting using the indicated antibodies.

For all other exogenously expressed epitope-tagged mammalian target proteins, cells were harvested after 48 h of transfection and lysed for 2 h at 4°C by rotating at 16 rpm in BC300 buffer (20 mM Tris-Cl [pH 8.0], 300 mM KCl, 2 mM EDTA, and 20% glycerol) containing 0.1% NP-40 and protease inhibitor cocktail. From the whole-cell extract, target proteins were pulled down using tag-specific-antibody-coated agarose beads. Subsequently, after overnight incubation, the beads were washed three times (10 min each at 4°C) using washing buffer containing BC300 plus 0.1% NP-40 to remove nonspecific interactors. Subsequently, bead-bound proteins were eluted in 1× SDS loading dye by incubating at 95°C for 10 min. The eluted proteins thus obtained were subsequently used for Western blotting.

Generation of stable cells.

In this study, we used two different stable cell lines expressing FLAG-HA-tagged FKBP5 and CDK9 proteins. Generation of a CDK9 stable cell line was already been reported in our earlier study (22). To generate a FLAG-HA-FKBP5-expressing stable line, we used the Flp-In method (Invitrogen) following the manufacturer’s protocol. Flp-In 293T cells were transfected with the appropriate pCND5-FRT-TO vector containing the FKBP5 open reading frame (ORF). The transfected cells were subjected to hygromycin (200 µg/mL) selection for several weeks to obtain the positive colonies. Individual colonies were picked up, expended, and further screened for the expression of the FLAG-HA-FKBP5 protein by Western blotting using FLAG tag-specific antibody.

Generation of stable knockdown cells.

Stable knockdown of FKBP5 in mammalian cell lines was obtained through lentivirus-mediated stable integration of target shRNA sequences against FKBP5 and CDK9. To produce specific lentiviruses, target sequences were first cloned in lentiviral pLKO.1 puro vector. Subsequently we cotransfected the shRNA construct along with psPAX2 (lentivirus packaging plasmid) and pMD2.G (lentivirus envelope plasmid) in 293T cells. Lentivirus particles were produced and collected after 72 to 96 h of transfection. Collected lentivirus particles were used with Polybrene (8 µg/mL) for transduction. Subsequently, transduced cells were selected using puromycin (3 μg/mL), and knockdown efficiency was measured at both RNA and protein levels.

Baculovirus expression-based reconstitution and purification.

To reconstitute and purify proteins or protein complex from insect cells, we used a baculoviral expression system. For the purification of expressed proteins/complexes, baculovirus particles were used directly to the 90% confluent Sf9 cells either alone or in combination, per experimental requirement. Cells were harvested after 48 h of infection. Cell lysate was prepared in BC300 buffer (20 mM Tris-Cl, 200 mM KCl, 2 mM EDTA, and 20% glycerol), containing 0.1% Nonidet P-40, 0.7 μL/mL β-mercaptoethanol, and 1× protease inhibitor cocktail for 2 h at 4°C. To get rid of cell debris, cell lysate was centrifuged at 12,000 rpm for 20 min at 4°C. Subsequently, immunoprecipitation was performed using epitope-tag-specific agarose beads to pull down protein-protein complex. After overnight binding, beads were washed three times (10 min each at 4°C) with washing buffer containing BC300 plus 0.1% NP-40. This bead-bound protein-protein complex was either used directly or eluted by competitive elution using epitope-tag-specific peptide.

Baculovirus expression-based interaction analysis.

For baculovirus expression-based interaction analysis, Sf9 cells were coinfected with the combination of viruses as indicated. Cells were harvested with phosphate-buffered saline (PBS) after 48 h of infection. Subsequently, cell lysate was prepared using BC300 buffer containing 0.1% Nonidet P-40, 0.7 μL/mL β-mercaptoethanol, and 1× protease inhibitor cocktail for 2 h at 4°C. The prepared lysate was subjected to immunoprecipitation overnight at 4°C using antibody-coated agarose beads. Protein-bound beads were washed three times (10 min each at 4°C) with washing buffer containing BC300 plus 0.1% NP-40. The bound proteins were eluted by incubating at 95°C for 10 min in 1× SDS loading dye. To identify the interactors, Western blot analyses were performed with specific antibodies.

In vitro protein-protein interaction assay.

For direct in vitro protein-protein interaction assay, we used purified recombinant target factors, as mentioned, either bound to agarose beads or added exogenously. Purified proteins were incubated in a binding buffer containing 20 mM Tris (pH 8.0), 20% glycerol, 2 mM EDTA, 150 mM KCl, 0.1% Nonidet P-40, and 20 to 80 ng/μL bovine serum albumin (BSA) and incubated for overnight at 4°C. Subsequently, protein-bound beads were washed three times (10 min each at 4°C) with binding buffer. The bound proteins were eluted by incubating at 95°C for 10 min in 1× SDS loading dye. To identify the interactors, Western blot analyses were performed with specific antibodies as indicated.

In vitro kinase assay.

In vitro kinase assays were performed to test the effect of FKBP5 on P-TEFb-mediated phosphorylation of its substrates Pol II CTD, DSIF complex, and NELF complex. Phosphorylation reactions were carried out in 1× kinase assay buffer (50 mM Tris-Cl [pH 8.0], 2 mM MgCl₂, and 500 μM ATP) at 30°C for 2 h. After completion of phosphorylation reactions, protein phosphorylation was assessed by Western blot analysis using specific antibodies (for GST-CTD phosphorylation) or by autoradiography (for NELF and DSIF complexes). For the purpose of phosphorylation assay of DSIF and NELF complexes, along with the 500 μM ATP, 0.2 μM γ-³²P-ATP was used additionally.

In vitro competitive binding assay.

For the in vitro competitive binding assay, we used purified bead-immobilized FLAG-CDK9 along with exogenously added GST-cyclin T1 and His-FKBP5. We used a constant amount (0.5 μg) of bait protein (immobilized FLAG-CDK9) along with a constant amount of one of the proteins (GST-cyclin T1/His-FKBP5), with the other competitor (His-FKBP5/GST-cyclin T1) being added in a gradient fashion. The binding reaction was carried out overnight at 4°C in a binding buffer containing 20 mM Tris (pH 8.0), 20% glycerol, 2 mM EDTA, 100 mM KCl, 0.1% Nonidet P-40, and 80 ng/μL BSA. After overnight incubation, the beads were washed in binding buffer rigorously. Bead-bound proteins were eluted by incubating at 95°C for 10 min in 1 × SDS loading dye. To identify the interactors, Western blots were performed with specific antibodies, as mentioned.

In vitro dissociation assay.

For the in vitro dissociation assay, we used purified bead-immobilized preformed P-TEFb complex composed of FLAG-CDK9 plus GST-cyclin T1. Both of these proteins were individually purified and assembled into the P-TEFb complex by adding them together in assembly buffer (20 mM Tris [pH 8.0], 20% glycerol, 2 mM EDTA, 300 mM KCl, and 0.1% Nonidet P-40) and incubating overnight at 4°C. Bead-bound FLAG-CDK9 was incubated with an excess amount of purified GST-cyclin T1 for fully saturating FLAG-CDK9 binding with GST-cyclin T1. The excess unbound GST-cyclin T1 was removed by extensive washing. This immobilized P-TEFb complex was then used in the reaction setup, along with increasing concentrations of purified FKBP5 protein, as mentioned, in reaction buffer (20 mM Tris [pH 8.0], 20% glycerol, 2 mM EDTA, 100 mM KCl, 0.1% Nonidet P-40, and 80 ng/μL BSA) and then incubated overnight at 4°C. The reaction mixtures were subjected to brief centrifugation, and the supernatant fraction was collected. The bead-bound proteins were washed twice with reaction buffer before eluting in 1× SDS loading dye by incubating them at 95°C for 10 min and were treated as bead-bound fraction. Both the supernatant and bead-bound fractions were loaded in parallel for comparing the amount of target proteins present in both fractions by Western blot analyses with specific antibodies, as mentioned. For immunoprecipitated P-TEFb complex, similar experimental approach was used with bead-bound P-TEFb complex immunoprecipitated from whole-cell extract from mammalian cells along with purified FKBP5 protein.

Dual luciferase assay.

To perform this assay, 50,000 cells of both scramble and stable FKBP5-KD cell lines were seeded into a 24-well dish. The next day, we transfected HIV1-LTR-luciferase (firefly) construct in three different concentrations along with equal amount of Renilla luciferase construct (as an internal control for transfection) in both of the cell lines. After 40 h of transfection, cells were harvested and proceeded for luciferase assay using the dual-luciferase assay kit from Promega and following the manufacturer’s protocol. The firefly luciferase data obtained were normalized with the internal control, Renilla luciferase.

qRT-PCR analysis for RNA expression.

To perform qRT-PCR analysis for mRNA expression of target genes, total mRNA extraction was done using TRIzol reagent (Invitrogen, Inc.) following the manufacturer’s protocol. Subsequently, 1 μg of extracted mRNA was used to synthesize cDNA using the Verso cDNA synthesis kit (Thermo Scientific) by following the manufacturer’s protocol. Synthesized cDNA was diluted 50× before being used for qRT-PCR analysis. qRT-PCR analysis was performed using iTaq Universal SYBR green supermix (Bio-Rad) and target gene-specific primers. qRT-PCR data were analyzed using CFX96 real-time PCR system software using GAPDH as an internal control for normalization.

Chromatin immunoprecipitation analysis.

ChIP analysis was performed essentially following a previously described protocol (37). Briefly, cells were cross-linked with 1% formaldehyde (Sigma) for 10 min at room temperature, followed by incubation with 125 mM glycine for 5 min to stop the cross-linking. The cross-linked cells were washed three times with ice-cold PBS. Nuclear extract was prepared using these cells for downstream immunoprecipitation analysis by resuspending the cells in ChIP lysis buffer (0.5% NP-40, 1% Triton X-100, 300 mM NaCl, 20 mM Tris [pH 7.5], 2 mM EDTA, and protease inhibitor cocktail) and incubating on ice for 30 min. After syringe passage 8 times through a 23-gauge needle, the lysate was spun down at 5,000 rpm for 10 min at 4°C. The obtained nuclear pellet was resuspended with ChIP sonication buffer (1% SDS, 50 mM Tris [pH 8.0], 10 mM EDTA, and 1× protease inhibitor cocktail). The resuspended nuclear pellet was subsequently sonicated at high efficiency using a Bioruptor UCD-200 instrument for 20 min. The sonicated samples were spun down at 15,000 rpm for 20 min. The obtained supernatant was precleared using 20 μL protein A-agarose beads in 65 μg of sonicated DNA sample for 30 min at 4°C. The precleared lysate was diluted 10 times using ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.1 mM EDTA, 20 mM Tris-Cl [pH 8.0], and 167 mM NaCl). Immunoprecipitation was carried out overnight at 4°C with 2 μg of target antibody for each immunoprecipitation sample, while IgG was used as a negative control. In a parallel reaction, protein G magnetic beads were blocked using 50 μg of salmon sperm DNA for overnight. The next day, 25 μL of preblocked protein G magnetic beads were further added with the immunoprecipitated sample and incubated for 1 h at 4°C. After a short spin at 3,000 rpm, magnetic beads bound to antibody-protein-DNA complex were successively washed twice with low-salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-Cl [pH 8.0], 150 mM NaCl, and protease inhibitor cocktail), high-salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-Cl [pH 8.0], 500 mM NaCl, and protease inhibitor cocktail), and lithium chloride buffer (0.5 M LiCl, 1% NP-40, 1% deoxycholate, 20 mM Tris-Cl [pH 8.0], and 1 mM EDTA). The immunoprecipitated protein-DNA complexes were eluted by incubating for 30 min with elution buffer (1% SDS and 0.1 M NaHCO₃) at room temperature. The eluted DNA-protein complex was incubated with 190 mM NaCl at 65°C for de-cross-linking. The remaining proteins were digested by using proteinase K for 45 min at 45°C. The eluted DNA was further purified using the Qiagen PCR purification kit following the manufacturer’s protocol. The purified eluted DNA was subsequently used in qRT-PCR analysis using primers specific to the target gene locus.

Cell proliferation assay.

To perform the cell proliferation assay, 6 × 10⁴ FKBP5 knockdown and control cells were seeded in a 6-well plate. Cell numbers were counted using a hemocytometer on the 3rd, 4th, and 5th days after seeding of cells. Scramble knockdown cells were used as control.

RNA-Seq analysis.

For the purpose of analysis of differential expression of genes between control (scramble) and FKBP5 knockdown cells, raw data were generated following the same protocol mentioned in our earlier study (37). The data obtained in the previously reported control sample (37) were used for reanalyzing the effect of FKBP5 knockdown, reported here, since all these experiments were performed at the same time. For differential gene expression analysis, the raw FASTQ sequences of all 4 samples (as obtained) were mapped to the Homo sapiens GRCh38 genome using STAR (v2.27.2b) to create BAM files (51, 52). The BAM files were then processed using the Rsamtools, Rsubread, and GenomicAlignments R packages, and abundance tables were created from the alignments (53, 54). The abundance tables were further normalized and differentially abundant genes were identified using the DESeq2 R package (55). Genes with a log₂ fold change (log₂FC) of 1.5-fold or more, having a P value of <0.05 and an adjusted P value of <0.05 were considered significantly differentiated gene expression. The heatmap of the selected genes was generated using counts-per-million normalization from raw counts by implementing the edgeR package, and subsequent visualization was done with “pheatmap” R function (56).

Gene ontology analysis.

Gene ontology (GO) analysis for the identification and classification of biological processes was performed as described previously (37). Functional classification of differentially regulated genes identified from the RNA-Seq data was done utilizing the Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.8 (57). GO terms with P values of >0.05 as determined by the EASE score, a modified Fisher’s exact P value test integrated into the DAVID workflow, were considered significant.

RNA expression analysis using TCGA data sets.

Gene expression data were obtained from The Cancer Genome Atlas (TCGA) database (data version 2016_01_28) stored as part of the Broad Institute of MIT and Harvard TCGA Government Data Analytics Center (GDAC) (www.gdac.broadinstitute.org). mRNA levels of indicated genes in 11 cancers were obtained from a TCGA database of 38 different cancer cohorts containing more than 14,000 sequencing samples. For easy data retrieval and analysis, the Firehose analysis pipeline developed by the Broad Institute itself was utilized through its Java-dependent web application FireBrowse (http://firebrowse.org/) (58). Cohort-specific mRNA abundance was represented as log₂ RNA-Seq by expectation maximization (RSEM) for greater accuracy in quantification (59). Statistical analysis was performed using a one-tailed Student’s t test, defining a P value of <0.05 as significant.

Data availability.

The full set of mass spectrometry data for identifying FLAG-HA-CDK9-associated proteins purified from mammalian 293T cells can be obtained from Mendeley Data at https://data.mendeley.com/datasets/sxndp9wpbx/1. Original raw images of Western blots used for making the figures shown in this study can also be accessed from Mendeley Data repository at https://data.mendeley.com/datasets/3c9vdxrf3n/1.