Kennedy’s Disease PHOSPHORYLATION OF THE POLYGLUTAMINE-EXPANDED FORM OF ANDROGEN RECEPTOR REGULATES ITS CLEVAGE BY CASPASE-3 AND ENHANCES CELL DEATH*

X-linked spinal and bulbar muscular atrophy is a degenerative disease affecting motor neurons that is caused by polyglutamine (polyQ) expansion within the androgen receptor (AR). The polyQ-expanded form of AR is cytotoxic to cells, and proteolytic cleavage enhances cell death. The intracellular signaling pathways activated and/or required for cell death induced by the expanded form of AR (AR112) are unknown. We found that AR regulates mitogen-activated protein kinase (MAP kinase) pathways and, therefore, hypothesized that these pathway(s) may be required for AR112-in-duced cell death. The polyQ expansion in AR activates three MAP kinase pathways, causing increasing levels of phosphorylation of p44/42, p38, and SAPK/JNK MAP kinase. Inhibitors of either the JNK or p38 pathways had no effect on AR112-induced cell death, suggesting they are not required for polyQ-induced cell death. Strikingly, the MEK1/2 inhibitor, U0126, which selec-tively inhibits the p44/42 MAP kinase pathway, reduces AR112-stimulated cell death. The inhibition of the MEK1/2 pathway correlates directly with a change in phosphorylation state of the androgen receptor. Mutation of the MAP kinase consensus phosphorylation site in AR at serine 514 blocked AR-induced cell death and the generation of caspase-3-derived cleavage products. We propose a mechanism by which phosphorylation at serine 514 of AR enhances the ability of caspase-3 the MEK1/2 phosphorylation of at Ser-514 does the of caspase-3 cleave

X-linked spinal and bulbar muscular atrophy is a degenerative disease affecting motor neurons that is caused by polyglutamine (polyQ) expansion within the androgen receptor (AR). The polyQ-expanded form of AR is cytotoxic to cells, and proteolytic cleavage enhances cell death. The intracellular signaling pathways activated and/or required for cell death induced by the expanded form of AR (AR112) are unknown. We found that AR regulates mitogen-activated protein kinase (MAP kinase) pathways and, therefore, hypothesized that these pathway(s) may be required for AR112-induced cell death. The polyQ expansion in AR activates three MAP kinase pathways, causing increasing levels of phosphorylation of p44/42, p38, and SAPK/JNK MAP kinase. Inhibitors of either the JNK or p38 pathways had no effect on AR112-induced cell death, suggesting they are not required for polyQ-induced cell death. Strikingly, the MEK1/2 inhibitor, U0126, which selectively inhibits the p44/42 MAP kinase pathway, reduces AR112-stimulated cell death. The inhibition of the MEK1/2 pathway correlates directly with a change in phosphorylation state of the androgen receptor. Mutation of the MAP kinase consensus phosphorylation site in AR at serine 514 blocked AR-induced cell death and the generation of caspase-3-derived cleavage products. We propose a mechanism by which phosphorylation at serine 514 of AR enhances the ability of caspase-3 to cleave AR and generate cytotoxic polyQ fragments.
Spinal and bulbar muscular atrophy (SBMA), 1 or Kennedy's disease, is an X-linked autosomal dominant degenerative disease of the motor neurons, characterized by progressive muscle atrophy and weakness in male patients (1). The disease is caused by a polyglutamine (polyQ) tract expansion within the transcriptional activation domain of androgen receptor (AR) (2). The normal number of CAG repeats is 11-36, whereas the mutant gene has up to 68 repeats. Presumably SBMA is caused by a gain-of-function mutation because severe testicular feminization patients where AR is fully inactivated do not have motor neuron disease (3,4). However, male patients with SBMA exhibit some symptoms of partial androgen insensitivity, including gynecomastia, small testicles, and feminization, suggesting some impairment of AR function (5). SBMA is one of a growing number of polyQ-repeat diseases, including Huntington's disease and the spinocerebellar ataxias, that are characterized by the death of specific neuronal subsets (6). It is particularly attractive to study the effect of a polyQ expansion in the context of AR because, unlike the other polyQ disease proteins, it has a well characterized function. AR is a transcription factor and a member of a large family of steroid hormone receptors (7,8).
Almost all the polyQ proteins identified to date are substrates for cell death proteases or caspases (9 -13). The expanded form of AR is toxic to cells in both cell culture systems and transgenic animal models (9,14). Recent studies both in Drosophila and in transgenic mouse models of SBMA, expressing full-length mutant protein, demonstrate that the addition of ligand accelerates the phenotypes in these models and truncated fragments are found in the affected tissue as well as postmortem SBMA brain tissue (15,16). We have shown previously that AR is a substrate for caspases, the disease form of AR is toxic to cells, and mutation of the caspase cleavage site in AR attenuates cell death (9). These observations along with other recent studies suggest that proteolytic cleavage by caspases may be important in SBMA and other polyQ diseases (17,18).
Because AR and other steroid receptors are highly dependent upon phosphorylation for function as transcription factors (19,20), we reasoned that a critical step in the production of proteolytic fragments might require phosphorylation. Further, the MAP kinase signal transduction pathway is involved in AR function and phosphorylation (21)(22)(23)(24). Therefore, we tested whether the polyQ expansion in AR activates three major MAP kinase pathways (p44/42, p38, and SAPK/JNK MAP kinases), whether these pathways were required for AR112-mediated cell death, and whether phosphorylation of AR112 was required for AR112mediated cell death. We found the p44/42 MAP kinase pathway is required for AR112 cell death and mutation of the MAP kinase consensus phosphorylation site in AR at serine 514 blocked ARinduced cell death and the generation of caspase-3-derived cleavage products. The results demonstrate the proteolytic cleavage of mutant AR is modulated by phosphorylation and demonstrate the MEK1/2 inhibitor, U0126, can be used to block AR112-induced cellular toxicity.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-HEK 293T cells were cultured as described previously (9). Transient transfections were performed using * This work was supported by National Institutes of Health Grant NS40251, the Huntington's Disease Society of America, the Hereditary Disease Foundation, and the Multiple Dystrophy Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cytotoxicity Assay-Cells were transiently transfected in 6-well dishes. After 48 h, cells were harvested by centrifugation at 500 ϫ g for 10 min. Cells were lysed (Apoalert kit lysis buffer, Clontech) for 10 min and assayed according to the manufacturer's instructions. Lysates (20 g) were resolved by SDS-PAGE, and Western blotting was performed using the AR N-20 and tubulin antibody to confirm equal expression. Lactate dehydrogenase assays were performed according to Ref. 25.
Subcellular Fractionation-293T cells were transiently transfected and harvested at 48 h by scraping, and subcellular fractionation was carried out as described previously (26). Nuclear fraction was verified by immunoblotting with anti-PARP antibody (1:500; Biomol).
In vitro Translation and Caspase-3 Cleavage-AR constructs were in vitro translated (IVT) as previously described (9). IVT reaction (5 l) was incubated with various concentrations of caspase-3 (gift from Guy Salvesen) for 4 h at 37°C. Cleavage products were resolved using 4 -12% BisTris precast gels (Invitrogen). Gels were dried and autoradiographed, and densitometry was performed using ChemiImager software (Alpha Innotech).

Expression of the PolyQ-expanded Form of AR Activates MAPK
Pathways in 293T Cells-The MAP kinase pathways are known to relay, integrate, and amplify signals from a number of physiological responses, including differentiation, proliferation, inflammation, and apoptosis. Recently, using gene chip analysis, we have found that AR regulates a number of genes involved in the MAP kinase pathway and the polyQ-expanded form of AR differentially regulates these genes. 2 We hypothesized that expression of the expanded form of AR, which induces cell death, would result in alterations in the MAP kinase pathway(s). To evaluate the effect of expression of AR12 or AR112 (12 or 112 glutamines, respectively) on the activation of the MAP kinase pathways, 293T cells were transiently transfected with AR constructs. In our cell culture model, expression of AR112 alone is sufficient to induce caspase activation and subsequent cell death. MAP kinase activation was determined by Western blot analysis with phospho-specific antibodies that recognize only the phosphorylated forms of the MAP kinases ( Fig. 1, A-C). AR112 leads to increased phosphorylation of p44/42 MAPK (Fig. 1A), SAPK/ JNK MAPK (Fig. 1B), and the p38 MAPK ( Fig. 1C) when compared with AR12 or vector control. Reprobing the Western blot confirmed equal expression of AR, total p44/42 MAPK, and ␤-tubulin ( Fig. 1, D and E). The lower band in the AR112 lane is a proteolytic fragment. These results suggest that activation of a MAPK pathway might play a role in expanded AR-induced cell death.
MEK1/2 Pathway Is Required for AR-induced Cytotoxicity-To investigate whether the cytotoxicity of AR112 is mediated by MAPK pathways, we treated 293T cells that had been transiently transfected with AR112 with MAP kinase inhibitors: the MEK1/2 inhibitor, U0126, the SAPK/JNK inhibitor, SP-600125, or the p38 MAPK inhibitor, SB 203580. Forty-eight hours after treatment, cytotoxicity was evaluated by monitoring caspase-3 activity (Fig. 2). Addition of the MEK1/2 inhibitor, U0126, dramatically reduced AR112-induced cytotoxicity, reducing it to levels of vector alone ( Fig. 2A). Treatment with either a JNK/SAPK inhibitor (SP-600125) or a p38 MAPK inhibitor (SB 203580) did not reduce the toxicity of AR112 ( Fig.  2A). Similar results were observed using a cell viability (WST) assay (data not shown). Interestingly, MEK1/2 inhibitor treatment also inhibits the cleavage of AR112 (Fig. 2B). Cleavage of AR has been correlated with cellular toxicity (9). As a control for kinase inhibition, Western blot analysis demonstrated that the inhibitor doses utilized led to a decrease in phosphorylation of the kinases themselves and of ATF-2, a downstream target of the JNK pathway (Fig. 2C). Inhibitor treatments did not affect the expression level of AR. It is possible that the SAPK/JNK and p38 MAPK (stress-activated) pathways are activated in a protective response to AR112 cytotoxicity; this explains why the inhibitors of these pathways do not decrease toxicity. Indeed, if these pathways are activated as a protective mechanism in response to cellular stress, inhibition would be expected to increase toxicity, as is observed here (Fig. 2A). Our results suggest that the MEK1/2 pathway is required for AR112-induced cell death.
Serine Phosphorylation of AR is PolyQ Repeat Length-dependent-Given the increased activation of the MAP kinases, which are serine/threonine kinases, we examined whether AR might be a downstream target of serine phosphorylation. We immunoprecipitated AR12, AR50, or AR112 from transiently transfected 293T cells and performed Western blot analysis. An antibody that recognizes phosphorylated serine with a proline immediately C-terminal (phospho-Ser-Pro) (Fig. 3A, right) was used to evaluate AR phosphorylation because only six serineproline sites are present in AR. Equivalent amounts of AR were pulled down in each immunoprecipitation (Fig. 3A, left panel). Increasing amounts of phosphorylated serine were found in AR with 50 or 112 repeats (Fig. 3A, right panel), although AR12 was not phosphorylated at this site. The nonspecific band that is recognized by the phosphoserine antibody (denoted by *) is not AR, because the AR antibody does not recognize this band in the vector control lane and this band runs at a molecular weight that is between the molecular weight of the AR12 and AR50 bands. Therefore, the expanded form of androgen receptor has increased serine phosphorylation, correlating with the increase in MAP kinase activation.
Phosphorylation of Serine 514 Is Required for AR112-induced Cytotoxicity-Several serine residues, including Ser-19, -210, -514, and -790, are phosphorylated in AR. Phosphorylation at these sites can lead to an increase, no change, or a decrease in AR transactivation activity (20,21,27). Particularly of interest is Ser-514, in the transcriptional activation domain (Fig. 3B), which is phosphorylated by MEK1/2 in a prostate cancer cell line (21), is in a MEK1/2 consensus se-quence that has a proline at the ϩ1 position (Fig. 3C), and would be recognized by the serine phospho-specific antibody utilized in Fig. 3A. Treatment of AR112-expressing cells with the MEK1/2 inhibitor U0126 reduced the phosphorylated form of the AR that is immunoreactive to the phospho-Ser-Pro antibody (Fig. 3D, upper panel). Treatment with DHT, which often modulates phosphorylation of AR, did not alter phosphorylation of AR immunoreactive to the phospho-Ser-Pro antibody (Fig. 3D, lower panel). Inhibition of phosphorylation of AR by U0126 suggests the serine phospho-specific antibody recognizes a site in AR phosphorylated by MEK1/2 or downstream targets of this pathway. Ser-28 and Thr-438 also lie in the transcriptional activation domain (Fig. 3B) and are adjacent to FXXLF motifs, which mediate interactions of AR with its transcriptional co-activators. These sites are consensus sites for p90RSK, a kinase that phosphorylates transcription factors and is a downstream target of MEK1/2 (28).
We used site-directed mutagenesis to eliminate these phosphorylation sites. Mutation of any of these sites eliminates the doublet that is seen in cells expressing AR112, suggesting that these sites affect post-translational modification of AR112 (Fig.  3E). Two isoforms of AR112 are reactive to the phospho-Ser-Pro antibody (Fig. 3F). Immunoreactivity of one of the isoforms of AR112 to the phospho-Ser-Pro antibody is eliminated by mutation of Ser-514 to Ala (Fig. 3F).
To investigate the effect of phosphorylation on AR112-induced cytotoxicity, we expressed AR112, AR112 (S28A), AR112 (T438A), and AR112 (S514A) in 293T cells and performed cytotoxicity assays. Transient expression of AR112 (S28A) and AR112 (T438A) increased (1.5-fold) caspase-3 activity (Fig. 4A). However, expression of AR112 (S514A) decreased cytotoxicity (2-fold) in 293T cells. Expression levels of AR112 and the mutants in lysates used for these assays were equivalent (Fig. 3E). Similar results were obtained using a lactate dehydrogenase assay to measure cell death (data not shown). Interestingly, both AR112 and AR112 S514A activate p44/42 MAP kinase pathway (data not shown), suggesting the polyQ expansion activates this pathway but kinase activation is not sufficient for cell death. To further evaluate whether the phosphorylation status of androgen receptor affected cellular cytotoxicity, we transiently transfected mouse motor neuron-neuroblastoma (MN) hybrid cells, a neuronal cell line directly relevant to SBMA with AR112 and AR112 (S514A). We found, by LDH assay (Fig. 4B) and by caspase activation assay (data not shown), that AR112 (S514A) was less toxic than AR112. Because we have previously reported that the caspase-resistant form of the polyQexpanded form of AR has reduced cellular toxicity, we compared AR112 (S514A) to AR112 (D154N). Interestingly, the decreased cytotoxicity of the AR112 (S514A) was similar to that observed for AR112 (D154N), a caspase-resistant form of AR that cannot generate N-terminal polyQ-containing caspasemediated cleavage products (Fig. 4B). Next we evaluated whether addition of DHT, which results in nuclear localization and transcriptional activation of AR, modulated cellular toxicity. Cellular toxicity of AR112 expressed in MN hybrid cells was enhanced (Fig. 4B, 1.3-fold, **, p Ͻ0.01). DHT did not modulate the toxicity of cells expressing AR112 (S514A) or AR112 (D154N). Expression levels of AR112 and the mutants in lysates used for these assays were equivalent (Fig.  4B). The effect of the Ser-514 mutant is consistent with the decrease in cytotoxicity seen upon treatment of cells with the MEK1/2 inhibitor and further suggests that the MEK1/2 pathway, and specifically MAPK phosphorylation of AR112, is required for cytotoxicity in our cell culture model. Therefore, we focused on the AR112 (S514A) mutant for further characterization.
Phosphorylation of Ser-514 Does Not Alter Localization of AR112-In many polyQ diseases, nuclear localization correlates with enhanced cell dysfunction or cell death (29 -31). Serine phosphorylation regulates nuclear translocation of proteins, and therefore we investigated whether the reduced cytotoxicity of the AR112 (S514A) mutant was due to altered subcellular localization. Cells transfected with AR112 and AR112 (S514A) were treated with dihydrotestosterone and fractionated into cytoplasmic and nuclear fractions. Western blot analysis demonstrates that the distribution between the cytoplasm and nucleus and translocation for the full-length AR112 and AR112 (S514A) were similar in response to DHT (Fig. 5A). Western blot analysis with a nuclear fraction-specific marker, PARP, was performed to verify complete fractionation (Fig. 5A). Our results suggest that Ser-514 phosphorylation does not alter translocation of AR112 and therefore the reduction in cytotoxicity of AR112 (S514A) is mediated by another mechanism.
Phosphorylation of Ser-514 Modulates Proteolytic Processing of AR112-We have previously shown that AR is cleaved by caspases and mutation of the caspase cleavage sites in AR reduces cellular toxicity (9). Therefore, we investigated whether phosphorylation of the Ser-514 site modulated proteolytic processing of AR. Western blot analysis of AR112 (S514A) demonstrates reduced proteolytic cleavage of this protein when compared with AR112 (Fig. 5A). Proteolytic processing of AR112 was enhanced in the presence of DHT (Fig. 5A), and mutation of the Ser-514 to alanine site diminished the production of these cleavage products. We previously mapped the 32-kDa product formed by caspase-3 cleavage of AR at aspartic acid 154 in the sequence DEDD (Fig. 5B) and demonstrated that generation of this polyQ AR fragment is toxic to cells because mutation at this site (D154N) greatly reduces cellular toxicity (9). This caspase-derived cleavage product is absent in the AR112 (S514A)-expressing cells (Fig. 5C). Mutation of S28A or T438A in AR had no influence on the generation of caspase-derived AR cleavage products (Fig. 5C).
We next investigated whether phosphorylation of AR112 at the Ser-514 site could affect cleavage in an in vitro translation assay with purified recombinant caspase. AR112 and AR112 (S514A) were in vitro translated and products were treated with caspase-3 (150 nM). As shown in Fig. 5D, full-length AR112 (S514A) is resistant to caspase-3 cleavage compared with AR112 and the 32-kDa caspase-3 cleavage product is not produced by the AR112 (S514A) mutant protein under these conditions.
To further investigate the effect of the S514A mutation on caspase-3 cleavage of AR, we measured the cleavage of AR112 and AR112 (S514A) in response to increasing doses of FIG. 3. Serine phosphorylation is polyQ repeat length-dependent in 293T cells. A, AR was immunoprecipitated and Western blot analysis was carried out with either AR N-terminal antibody (left) or phosphoserine-specific antibody 16B4 (proline consensus phosphoserine site, phospho-Ser-Pro, right). * indicates a nonspecific immunoreactive band. B, possible serine phosphorylation sites in AR transcriptional activation domain (TAD). C, the MAP kinase consensus sequence is conserved in human (h), mouse (m), and rat (r) AR. D, Western blot of immunoprecipitated AR112 treated with U0126 or 1 M dihydrotestosterone (DHT) probed with phosphoserine-specific antibody 16B4. E, mutation of serine phosphorylation sites in AR results in reduced isoforms when compared with AR112. Western blot was probed with AR N-terminal antibody. F, mutation of serine phosphorylation site 514 in AR results in the elimination of one isoform of AR112 immunoreactive to phosphoserine-specific antibody 16B4. AR112 was expressed in 293T cells in the absence of testosterone unless otherwise indicated.
caspase-3. Generation of half of the maximal amount of 32-kDa cleavage product of AR112 occurs at a caspase-3 concentration of 150 nM (Fig. 5E). However, the half-maximal amount of 32-kDa cleavage product of AR112 (S514A) occurs at a caspase-3 concentration of 500 nM. Furthermore, the amount of product formed from caspase-3 cleavage of AR112 (S514A) leveled out at only 40% of the amount that was generated by wild type AR112 cleavage, even at a caspase-3 concentration of 6 M (Fig. 5E).
Ligand Dependence of Phosphorylation and Proteolytic Cleavage of AR-It has recently been reported that addition of testosterone accelerates the disease onset and progression in mouse and Drosophila models of SBMA (15,16,32) and proteolytic cleavage may be involved. Therefore, we evaluated whether the proteolytic cleavage or phosphorylation of AR112 at Ser-514 was ligand-dependent because both these protein modifications enhance cell death (9,16). As shown in Fig. 5F, proteolytic cleavage of AR is both polyQ-and ligand-dependent (compare AR12 versus AR50 versus AR112, Ϫ/ϩ DHT). This correlates with the enhanced cytotoxicity of AR112 in the presence of DHT shown in Fig. 4B. As described previously, serine phosphorylation of AR112 at amino acid 514 was not enhanced by addition of DHT (Fig. 4D). Therefore, phosphorylation of AR112 at Ser-514 is polyQ repeat length-but not ligand-dependent. Cleavage of AR is modulated by polyQ repeat length and ligand. DISCUSSION The major findings of our studies are (i) AR112 leads to at least a 2-fold increase in phospho-p44/42, -JNK, and -p38; (ii) the MEK1/2 inhibitor, U0126, suppresses AR112-induced caspase activation and cell death; (iii) the phosphorylation of AR is repeat-dependent with increasing phosphorylation occurring with increasing polyQ length; (iv) mutation in the putative Inset shows immunoblot using anti-PARP antibody (nuclear marker) to confirm fractionation. A longer exposure was used to visualize cleavage products. B, mutation of serine 514 to alanine reduces proteolytic processing of AR112 and production of toxic N-terminal fragments of AR112 in 293T cell lysates. Schematic representation of AR112 caspase cleavage and phosphorylation sites. C, 293T cells were transfected with AR112, AR112 (S28A), AR112 (T438A), and AR112 (S514A) in the absence of testosterone. Cells were harvested after 48 h. Western blot analysis was carried out with AR N-terminal antibody. D, reduced proteolytic cleavage of in vitro translation of AR112 (S514A) with caspase-3 when compared with AR112. The production of the previously mapped caspase-3-derived 32-kDa AR N-terminal cleavage product is reduced. E, kinetic evaluation of AR112 and AR112 (S514A) with recombinant caspase-3. Varying concentrations of caspase-3 were used to cleaved 35 S-radiolabeled AR112 or AR112 (S514A) for 4 h at 30°C. F, proteolytic processing of AR is repeat length-and ligand-dependent. 293T cells were transfected with DNA (10 g) in 10-cm dishes, treated with 1 M DHT, and harvested after 48 h. Western blot analysis was carried out with AR N-terminal antibody. Note: in our fractionation, AR expression appears higher in the cytoplasmic fraction in the presence of DHT. This is likely because of a decrease in membrane-associated AR because total expression of AR shown in panel F is not changed in the presence or absence of DHT. It should be noted that the level of the 32-kDa fragment may appear low because of loss of insoluble fragment in the stacking gel.
proline-directed phosphorylation site, Ser-514, blocks AR gel motility and reactivity to serine phospho-specific antibody; and (v) mutation of Ser-514 in AR112 blocks AR-induced cell death, and the reduction in cellular toxicity is similar to that found for AR112 D154N, a form of AR resistant to caspase cleavage. Therefore, our studies demonstrate that phosphorylation of AR at Ser-514 promotes the cytotoxicity of the polyQ-expanded form of AR and addition of the specific MEK1/2 inhibitor, U0126, reduces cellular toxicity mediated by the polyQexpanded form of AR and generation of AR fragments. The phosphorylation of AR at Ser-514 does not alter the subcellular localization of AR, and previous studies have demonstrated that phosphorylation at this site does not alter the transcriptional activity of AR (21). Our results indicate that the expansion of AR induces ERK activation, which in turn phosphorylates AR112. Phosphorylation at Ser-514 modulates the ability of caspase-3 to cleave AR and produce cytotoxic fragments in either MN hybrid cells or HEK 293T cells.
It has been suggested that proteolytic cleavage of mutant protein may play an important role in the pathophysiology of SBMA as well as other polyQ diseases (33). We have previously shown that the toxicity of caspase-resistant expanded AR was markedly reduced in transfected motor neuronal cells (9). Our results, taken together, suggest a mechanism for AR-induced cytotoxicity. The phosphorylation of AR by MAPK is repeat length-dependent. The expansion in AR induces a conformational change (16,32) that activates MEK1/2, and this leads to enhanced serine phosphorylation of the mutant protein. Increased MAPK phosphorylation of the expanded form of AR may result in a further conformational change in AR that leads to enhanced proteolytic cleavage at the caspase cleavage site. Therefore, MAPK phosphorylation at Ser-514 of AR enhances the ability of caspase-3 to cleave AR and generate N-terminal fragments containing the polyQ repeat that are toxic to cells.
Both the phosphorylation site at 514 and the caspase cleavage site at amino acid 154 of AR are in the transcriptional activation domain but distal to each other. Modulation of conformation/function at sites distal to a phosphorylation site is not uncommon. C-terminal phosphorylation of Src kinase decreases ligand binding in the SH3 domain by inducing a conformational change in the N terminus of the protein (34). Our results do not exclude the possibility that MAPK phosphorylation of Ser-514 may mediate phosphorylation events at sites other than Ser-514 that may be important for promoting toxicity.
One particularly interesting finding is that proteolytic cleavage and toxicity is enhanced in the presence of DHT. Recent studies both in Drosophila and in transgenic mouse models of SBMA demonstrate that the addition of ligand accelerates the phenotypes in these models and truncated fragments are found in the affected tissue (15,16,32). Our results suggest that the toxic gain-of-function could be further defined as an increase in serine phosphorylation of mutant AR as a function of polyQ tract. This would lead to increased ligand-and repeat-dependent proteolytic cleavage of AR and correlates well with the results found in Drosophila and mouse SBMA models.
MAP kinase pathways exert multiple effects on cell survival and death, cell proliferation as well as stress response (35)(36)(37). We found that the polyQ-expanded form of AR increased the activation of JNK/SAPK, p38, and p44/42 MAP kinase pathways. Each of these pathways plays a role in cell death and survival. We found that only the p44/42 MAP kinase pathway was required for AR112-mediated cell death. Although we demonstrate that removal of the phosphorylation site in AR at Ser-514 reverses the cytotoxic effect of polyQ expansion, the inhibition of the p44/42 MAP kinase pathway with U0126 is likely to be protective through blocking the action of a number of cellular proapoptotic proteins that require phosphorylation to promote cell death. Inhibition of the p44/42 MAP kinase pathway protects against hippocampal neuronal cell death (38), although the mechanism is not known.
Our findings demonstrate that phosphorylation of AR enhances cellular cytotoxicity and the production of cytotoxic fragments. Recent work on SCA1 transgenic mice demonstrated that phosphorylation of Ser-776 of mutant ataxin-1 was critical for the formation of nuclear aggregates, interaction with 14 -3-3, and neurodegeneration (39,40). Phosphorylation of Htt by Akt in response to IGF-1 inhibits cellular cytotoxicity of Htt. In this case, phosphorylation was protective and did not alter the ability of Htt to be proteolytically cleaved by caspases or calpains (41). Phosphorylation of Htt more likely mediates protein-protein interactions that enhance cell survival. Phosphorylation has also been shown to play an important role in a number of neurodegenerative diseases, particularly Alzheimer's disease (42)(43)(44). It is clear that phosphorylation can modulate polyQ-dependent cytotoxicity by a number of different mechanisms. Future research will be directed at evaluating whether the repeat-dependent phosphorylation found for AR is a common feature of all polyQ diseases.