Structural and biochemical insights into FKBP51 as a Hsp90 co‐chaperone

1. INTRODUCTION

FKBP51 (encoded by the fkbp5 gene) was first identified as a protein component of Hsp90‐progesterone receptor complexes with an apparent molecular mass of 51 kDa. ¹ , ² Subsequent studies firmly established its interaction with Hsp90 and its role in steroid hormone receptor (SHR) signaling. ³ FKBP51 has three domains, ⁴ which mediate Hsp90 binding (tetratricopeptide repeat [TPR] and eH7), have scaffolding function (FK2), and display peptidyl‐prolyl‐isomerase activity and affinity to FK506 and analogs thereof (FK1, see Figure 1).

Interest in FKBP51 increased substantially when it became clear that FKBP51 is genetically associated in humans with endophenotypes that are thought to be linked to stress‐related diseases such as depression and posttraumatic stress disorder. ⁵ Parallel studies in FKBP51‐knockout mice clearly confirmed the role of FKBP51 as a key regulator of the mammalian stress response. ⁶ After the discovery of the first selective and in vivo applicable FKBP51 inhibitors, ⁷ these findings were largely replicated pharmacologically, suggesting FKBP51 as a potential novel drug target. ⁸

Key pending issues are the biochemical underpinnings, how FKBP51 contributes to pathobiological alteration at the molecular level. What are the relevant direct protein interaction partners and how does FKBP51 act on them? In this viewpoint, we focus on FKBP51 interactions that are postulated to be Hsp90‐dependent. Potential FKBP51 interaction partners on a broader scale were reviewed by Hähle et al. ⁵ Although the focus of this review is on FKBP51, we repeatedly will discuss also FKBP52 as its closest homolog, as their functions are often intertwined and the interplay between FKBP51 and FKBP52 is key to fully understand the function of either of these proteins.

2. HSP90‐DEPENDENT INTERACTION PARTNERS

A number of interaction partners have been described for FKBP51, usually based on co‐immunoprecipitation experiments. Several of these interactions have suggested to be Hsp90‐mediated. Here we will exemplarily discuss Tau‐ and cyclin‐dependent kinases (CDKs). The interaction with SHRs will be discussed in more detail in the next section.

FKBP51 was proposed to regulate the Alzheimer's disease‐related protein Tau. ⁹ , ¹⁰ , ¹¹ Jinwal et al. ⁹  suggested an inhibitory role of FKBP51 on Tau dephosphorylation leading to increase in Tau stabilization, whereas Blair et al. ¹⁰ proposed that FKBP51 shifts the folding of Tau to amorphous Tau aggregates. Contradictory, FKBP51 was shown to have both inhibitory ¹²  and enhancing effects on Tau phosphorylation levels. ⁹ Differences in the observed effect of FKBP51 on Tau phosphorylation could be explained by the different phosphorylation residues that were investigated.

FKBP51 was also found to associate with multiple CDKs, which play important roles in cell cycle progression. The best studied example is CDK4, which was proposed to directly bind to FKBP51 in co‐immunoprecipitation assays and pull‐down experiments with purified proteins. ¹³ , ¹⁴ Ruiz‐Estevez et al. ¹⁴ showed a delay in CDK4‐dependent myoblast differentiation after FKBP51 knockdown based on staining experiments in primary myoblasts. They suggested an important role of the Peptidyl‐Prolyl‐Isomerase (PPIase) activity based on mutant studies and hypothesized that FKBP51 retains CDK4 in the Hsp90 complex, priming it for dephosphorylation and further processing, resulting in continuation of differentiation. ¹⁴ In addition to CDK4, there is also evidence that CDK5 ¹⁵  interact with FKBP51 and FKBP52. Gassen et al. ¹² , ¹⁵ could show contrasting effects of FKBP51 overexpression on the kinase activity of CDK5 on two different substrates, both leading to a decrease in activity. The role of Hsp90 in this context remains to be elucidated.

Additionally, Mall et al. ¹⁶ showed in co‐immunoprecipitation assays combined with deletion mutants that the TPR domain of FKBP51 is essential for an interaction with CDK9. The authors claim a crucial role of FKBP51 in the regulation of proliferative genes via inhibition of CDK9 kinase activity. ¹⁶

3. FKBP51 AS A REGULATOR OF SHRS

Among all proposed interaction partners of FKBP51, the regulation of SHRs is best characterized. SHR undergo a complex maturation procedure orchestrated by the Hsp90 chaperone machinery as part of their activation mechanism. Unliganded SHRs reside in the cytosol predominantly in complex with Hsp90, which may help loading of steroid hormones to the hydrophobic steroid hormone‐binding site. Steroid binding triggers conformational changes that are thought to expose nuclear localization signals, allow recruitment of cofactors, and ultimately enable SHRs to act as transcriptions factors. FKBP51/52 are thought to bind to Hsp90 in late stages of the maturation process via their TPR domain. ¹⁷ , ¹⁸ It is generally accepted that the two immunophilins display antagonizing behavior on the glucocorticoid, progesterone (and, to a lesser extent, the mineralocorticoid) receptor, with FKBP51‐inhibiting and FKBP52‐potentiating receptor activity (Figure 2).

In contrast to other SHRs, it remains ambiguous whether FKBP51 has a suppressive or stimulatory role in androgen‐driven processes. In prostate cancer cells, FKBP51 was repeatedly described as a positive factor for androgen receptor (AR). ¹⁹ , ²⁰ , ²¹ , ²² However, in other mammalian cells (e.g., 52KOMEF cells), no activation of AR was observed, ²³ , ²⁴ , ²⁵ whereas FKBP52 did enhance AR activity in reporter gene assays. The underpinnings for this apparent cell context‐dependence are not clear.

Recently, Maeda et al. ²² demonstrated that FKBP51 and FKBP52 both promote AR activation, as defined by an AR dimerization assay, despite both having no effect on AR levels and localization. Intriguingly, they found that the selective FKBP51 inhibitor SAFit2 can attenuate antiproliferative effects on 22Rv1 cells to the same extent as a FKBP51 knockout. At the concentration used, SAFit2 can be expected to occupy intracellular FKBP51 ²⁶ but an additional blockade of FKBP52 cannot be excluded anymore. As a FKBP52 knockout could be fully rescued by overexpression of a PPIase‐dead FKBP52 mutant, Maeda et al. ²² concluded that the PPIase activity of FKBP51 but not FKBP52 leads to AR activation. For the specific F67Y mutant used by Maeda et al., ²² it has to be kept in mind that F67 and, likely even more so, Y67 can flip out of the PPIase‐binding pocket ⁷ , ²⁷ to adopt a conformation that was shown to be functional in a neurite outgrowth assay. ⁷ Taken together, FKBP51 and FKBP52 have emerged in recent years as potent potentiators of AR signaling in prostate cancer cells; however, the exact mode of action in AR signaling remains elusive.

For the estrogen receptor (ER), the impact of Hsp90‐dependent interaction has been investigated by Schülke et al. ²⁴ They found ER activity in reporter gene assays only to be moderately affected by Hsp90 inhibition. Additionally, they also describe the ER to be the SHR least influenced by TPR domain proteins. Yet, FKBP51 (and FKBP52) have both been described to be associated with the ER. ²⁴ This could be physiologically relevant, as FKBP52 has been shown to be elevated in ER‐positive breast cancer cell lines and FKBP52 overexpression correlates with poor survival rate in patients. ²⁸ , ²⁹ The latest study on the influence of FKBP51 and FKBP52 on ER by Habara et al. ³⁰ found FKBP52 to be a positive regulator of ERα. They further observed FKBP51 to be a negative regulator of ERα, similar to the well‐established antagonistic effects on the glucocorticoid receptor (GR). ³⁰ FKBP52 was described to have a stabilizing effect on ERα in cycloheximide chase experiments, which was shown to be impaired by a TPR‐dead mutant and an enzymatically nonactive PPIase‐dead mutant. The interaction of both FKBP51 and FKBP52 with ERα was shown to be impaired for the TPR‐dead but not by the PPIase‐dead mutant. This suggests that PPIase function of FKBP51 and FKBP52 is not important for binding to the SHR complex but needed for the stabilizing function described for FKBP52.

Among the SHRs, FKBP51's role in inhibiting GR activity, as well as the antagonistic effect by FKBP52 and the involvement of Hsp90, is experimentally best supported. The key question is the mechanistic basis for the antagonistic regulation of FKBP51 and FKBP52 on GR. The positive regulatory effect of FKBP52 on the GR has been repeatedly associated with a recruitment of dynein and enhancement of nuclear translocation of the GR. ³¹ , ³² , ³³ , ³⁴ Generally, the PPIase activity of FKBP51 is reported to be dispensable for GR regulation, although it is difficult to unambiguously demonstrate the involvement or lack thereof of regulated proline isomerization (also see below).

4. STRUCTURAL STUDIES

Another central question regarding the role of FKBP51 (and FKBP52) is the timing as well as the trigger of their entry and exit into and out of GR maturation complexes.

Ebong et al. ³⁵ used native mass spectrometry to structurally investigate in vitro complex formation of Hsp90/GR heterocomplexes including Hsp70 and the cochaperones FKBP51, FKBP52, Hop, Hsp40, and p23. In general, the GR ligand‐binding domain (LBD)‐containing complexes were found to be more stable in the absence of immunophilins. Complexes resembling early stages of the chaperoning cycle—defined by the absence of p23—showed FKBP51 and FKBP52 to be interchangeable. Interestingly, only FKBP51 and not FKBP52 was found to form a stable complex with p23, Hsp90, and the GR. FKBP52 was instead associated with p23 release and the formation of a complex Hsp90/GR/FKBP52 referred to as nuclear transfer complex. The data supported a scaffolding function by which FKBP51 stabilizes p23‐containing complexes, thereby stalling GR activation. Unfortunately, these experiments were performed in the absence of GR ligands. It is therefore unclear how the observed complexes change upon GR activation, as has been suggested in cellular experiments. Davies et al. ³² reported FKBP52 to be preferred after ligand binding over FKBP51 in GR complexes, which would be in line with the theory of a nuclear transfer complex containing FKBP52. An alternative route for translocalization proposes immediate dimerization of GR upon hormone binding in the cytosol. ³⁶

Recent biochemical studies further improved our structural understanding of FKBP51 binding to Hsp90. Oroz et al. ³⁷ used nuclear magnetic resonance studies together with chemical crosslinking to analyze the Hsp90/FKBP51 interaction interface. This showed that FKBP51 can interact with Hsp90 in the open conformation in a head‐to‐head orientation at higher concentrations. This could be driven by the MEEVD^Hsp90‐TPR^FKBP51 interaction alone, which would allow for a (Hsp90)₂(FKBP51)₂ stoichiometry as observed by Oroz et al. ³⁷ However, the open conformation proposed by Oroz et al. ³⁷ does not explain or is even at odds with the repeatedly observed preference of FKBP51 for p23‐containing Hsp90 complexes in cells.

The binding of FKBP51 to the closed conformation of Hsp90 has been structurally characterized in detail by a recent cryogenic electron microscopy structure of the Hsp90/FKBP51/p23 ³⁸ (Figure 3). Lee et al. ³⁸ observed that purified FKBP51 preferentially binds the closed, ATP‐bound Hsp90 dimer and forms a complex with an asymmetrical 2:1 stoichiometry, which is further stabilized by p23 binding. A key finding of the Hsp90/FKBP51/p23 structure is the binding motif of the extended C‐terminal helix (H7e^FKBP51) to a groove at the interface between two C‐terminal domains of Hsp90 (CTD^Hsp90) (Figure 3). The formation of this cleft was suggested to be coupled to the closed state of Hsp90, which would explain the preference for p23‐containing Hsp90 complexes. The CTD^Hsp90/H7e^FKBP51 interaction further suggests a preference for a (Hsp90)₂(FKBP51) stoichiometry and it beautifully explains earlier work by Smith and colleagues, ¹⁸ which indicated an important but incompletely understood role of the H7e for binding to Hsp90. The structure by Lee et al. ³⁸ positions the FK1 and FK2 domains of FKBP51 in an asymmetric manner close to the client binding channel at the core of Hsp90. ³⁹ However, in the absence of clients, the observed orientations of the FK1 and FK2 domains have to be regarded with caution. A direct overlay with the Hsp90/p23/LBD^GR structure ³⁹ (Figure 3) and the Hsp90/Cdk4/Cdc37 structure ⁴⁰ show massive clashes of these clients (LBD^GR or Cdk4) with the FK1 domain as positioned in the Hsp90/FKBP51/p23 complex. ³⁸ The FK1 and FK2 domains of FKBP51 and perhaps also the Hsp90 clients have to undergo substantial movements to accommodate each other in ternary complexes.

The Hsp90/FKBP51/p23 structure by Lee et al. ³⁸ represents a major advancement in our understanding how FKBP51 binds to Hsp90. However, it does not explain how FKBP51 interacts with clients (see above), how it suppresses GR activation, and most importantly, why its closest homolog FKBP52 has the opposite effect.

Solving structures of a client bound to the Hsp90/FKBP51/p23 complex as well as similar structures with FKBP52 might be necessary to fully understand the role of the FKBP51/FKBP52 cochaperones.

5. CONCLUSIONS

The recent finding discussed above provide intriguing new insight to our understanding of the molecular role of FKBP51. However, many challenges still remain. A key limitation is the unclear status of many suggested FKBP51 interaction partners. FKBP51 appears to be a rather promiscuous protein poised for protein‐proteins interactions (alone or mediated via Hsp90). Not all of the reported protein‐proteins interaction partners may be functionally relevant and many of the proposed FKBP51 interaction partners remain to be independently confirmed.

Rigorous mechanistic characterization of FKBP51, however, is far from trivial. FKBP51 may act only in the context of larger, transient complexes—as in the case of steroid hormone regulation. To the best of our knowledge, it was not possible so far to rigorously recapitulate the inhibitory effect of FKBP51 in a biochemical defined system, that is, to reconstitute these functions with purified proteins.

An important issue is the role of the PPIase activity and of the FK506‐binding site. This has therapeutic implications as these functions can be pharmacologically addressed with drug‐like compounds, whereas other functions currently cannot. Towards this end, PPIase‐deficient mutants are repeatedly used. While these experiments are important, they have to be interpreted with caution. Mutants devoid of PPIase activity in peptide substrate‐based assays may still retain PPIase activity towards peptidyl‐prolyl‐bonds in the context of a full protein. Moreover, the FK506‐binding pocket might act as a binding interface—as observed for FKBP12 and the ryanodine receptor 2 (RyR 2) ⁴¹ —and potential PPIase dead mutants might impair the interaction partner binding just as well as PPIase activity.

To further understand the molecular mechanism underlying FKBP51 biology, a rigorous characterization of its interaction with partner proteins at multiple layers of complexity is needed. As reductionist biochemical approaches can be misleading due to artefactual conditions, it is crucial to confirm biochemical findings in an authentic cellular context. Vice versa, it is crucial to validate protein interactions partners identified in cellular experiments in biochemically defined and functionally relevant assays.

AUTHOR CONTRIBUTIONS

Asat Baischew, Sarah Engel, Thomas M. Geiger, Martha C. Taubert, and Felix Hausch wrote the review and it was approved by all authors.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Baischew A, Engel S, Geiger TM, Taubert MC, Hausch F. Structural and biochemical insights into FKBP51 as a Hsp90 co‐chaperone. J Cell Biochem. 2024;125:e30384. 10.1002/jcb.30384

DATA AVAILABILITY STATEMENT

Data sharing not applicable to this article, as no data sets were generated or analyzed during the current study.