Control of Alzheimer's Amyloid Beta Toxicity by the High Molecular Weight Immunophilin FKBP52 and Copper Homeostasis in Drosophila

FK506 binding proteins (FKBPs), also called immunophilins, are prolyl-isomerases (PPIases) that participate in a wide variety of cellular functions including hormone signaling and protein folding. Recent studies indicate that proteins that contain PPIase activity can also alter the processing of Alzheimer's Amyloid Precursor Protein (APP). Originally identified in hematopoietic cells, FKBP52 is much more abundantly expressed in neurons, including the hippocampus, frontal cortex, and basal ganglia. Given the fact that the high molecular weight immunophilin FKBP52 is highly expressed in CNS regions susceptible to Alzheimer's, we investigated its role in Aβ toxicity. Towards this goal, we generated Aβ transgenic Drosophila that harbor gain of function or loss of function mutations of FKBP52. FKBP52 overexpression reduced the toxicity of Aβ and increased lifespan in Aβ flies, whereas loss of function of FKBP52 exacerbated these Aβ phenotypes. Interestingly, the Aβ pathology was enhanced by mutations in the copper transporters Atox1, which interacts with FKBP52, and Ctr1A and was suppressed in FKBP52 mutant flies raised on a copper chelator diet. Using mammalian cultures, we show that FKBP52 (−/−) cells have increased intracellular copper and higher levels of Aβ. This effect is reversed by reconstitution of FKBP52. Finally, we also found that FKBP52 formed stable complexes with APP through its FK506 interacting domain. Taken together, these studies identify a novel role for FKBP52 in modulating toxicity of Aβ peptides.


Introduction
FKBP52 is a high molecular weight FK506-binding immunophilin, possessing peptidyl-prolyl isomerase (PPIase) activity. It was first identified as a component of steroid hormone receptor heterocomplexes [1] and recently shown to regulate the nuclear localization of the glucocorticoid receptor [2]. It is widely expressed in mammalian tissues including the brain [3] and immunophilins FKBP12 and FKBP52 are up-regulated in regenerating neurons suggesting that they may play a protective or regenerative role following injury [4-7, reviewed in 8].
Immunophilins have also been associated with the processing of Alzheimer's Amyloid Precursor Protein (APP), which can be processed in an amyloidogenic or non-amyloidogenic manner. Overexpression of Pin1, a member of the parvulin family of immunophilins [9], reduces Ab and knockout of Pin1 increases Ab production in Alzheimer's disease brains, through the isomerization of the cytoplasmic domain of APP at a phosphorylated Thr 668 /Pro motif [10]. In addition, the APP intracellular domain (AICD) interacts with the peptidyl prolyl isomerase domain of the smaller immunophilin FKBP12 [11].
We have previously explored the existence of additional FKBP52-interacting cellular factors in neuronal cells and found that FKBP52 interacts with Atox1, a metallochaperone for copper [12]. In these studies, expression of FKBP52 in mammalian cells caused lower levels of intracellular copper, suggesting that FKBP52 facilitates copper efflux [12]. Metal dys-homeostasis is instrumental in the pathology of Alzheimer's disease and copper interactions with APP and Ab, both of which contain copperbinding sites, have been widely documented and implicated in the disease [13-14; reviewed in 15]. Although the direct interaction of copper and Ab is believed to be important for the aggregation and toxicity of the peptide, the copper/Ab interactions in vivo are complex and multifactorial. In particular, disturbances in both the intracellular compartmentalization of copper as well as in its extracellularly released forms may contribute to Ab production and toxicity [16]. Additionally, the interaction of APP with copper has been shown to alter levels of Ab in transgenic mice [17][18], although it is not clear whether this is a function of reduced Ab production or enhanced clearance of the peptide.
Given the importance of copper homeostasis in Alzheimer's disease pathology and our findings and published studies that immunophilins may participate in both of these processes, we examined if FKBP52 might affect Alzheimer's-related processes. To test this, we explored the function of FKBP52 using Drosophila genetics and found that mutations in FKBP52 exacerbate Ab toxicity while transgenic flies that overexpress wild type FKBP52 decrease Ab toxicity. The effects on Ab phenotypes correlated with altered levels of the peptide, suggesting that FKBP52 may affect Ab turnover. We also provide genetic and biochemical evidence that these effects of FKBP52 can be modulated by altering copper homeostasis during development. Finally, we also provide evidence that FKBP52 binds APP in mammalian cells and alters levels of Ab. Taken together, our data identify a novel role for FKBP52 in Alzheimer's disease, and suggests that this high molecular weight immunophilin acts on multiple aspects of Ab metabolism and toxicity.

Drosophila Strains, Rearing and Phenotypic Analysis
All flies were kept on yeast-containing media and were raised at 25uC or 29uC. dFKBP59 and Atox1 mutations were obtained from the Bloomington Drosophila Stock center (http://flystocks.bio. indiana.edu). UAS-Ctr1A flies are described in [19] and do not alter eye morphology when overexpressed (Fig. S1J). Based on information from Flybase [20] and our own analyses, dFKBP59 c01413 , dFKBP59 k00424 and dFKBP59 k09010 are viable lossof-function mutants and have no effect on the morphology of the eye ( Fig. 1A and Fig. S1E-F). Knock-out mutations of the mouse dFKBP52 gene are viable as well. The fact that three different insertional mutants in dFKBP59 cause the same phenotype in Ab flies supports the involvement of this gene in Ab toxicity. dFKBP59 EY03538 causes 2.9-fold up-regulation of the transcript in the presence of Gal4 protein (Fig. S2A) and does not alter eye morphology when overexpressed (Fig. S1K). Atox1 e01272 , Atox1 EY15780 and Atox1 f00729 are also loss of function mutations (Flybase, [20]), which do not affect eye morphology (Fig. S1G-I). Generation of the GMR-Ab42 and UAS-Ab42 flies is described in [21]. The UAS-Ab42 flies used in the lifespan analysis carry additional copies of the transgene, generated by re-mobilization of the original insertion. Additional production of Ab42 peptides accounts for their dramatically reduced lifespan. Expression of the Ab42 transgene in these flies was induced with the elavGal4 driver strain, which confers pan-neuronal expression [22]. Life-span monitoring was performed as described in [21] and Kaplan-Meier analysis was used to estimate lifespan probability for each genotype and feeding regimen. Time to event was defined as the lifespan for each fly. P-values were calculated using the log rank test on Stata Statistical Software, version 10 (StataCorp. 2007 College Station, TX: StataCorp LP). For the eye phenotype, flies were scored after aging to ,15-20 days old. The Ab42 effects are quantified by the severity of the rough eye phenotype, as previously [23]. When modifier mutations are co-expressed in Ab42-expressing flies, the distribution of phenotypes is shifted to more mild phenotypes in the case of a suppressor, or more severe phenotypes in a case of an enhancer (Fig. S1A-C and Table 1). ,100 flies were scored per genotype and each experiment was repeated 2 or 3 times and evaluated by a two-tail, two sample-equal variance Student's T-test.

Cu Feeding and Measurements
Males and female flies of the appropriate genotypes were put on food supplemented with CuCl 2 or the heavy metal chelator, bathocuproinedisulfonic acid (BCS; Sigma-Aldrich, St Louis, MO) as indicated. Progeny grew and were aged on the supplemented food until collected for analysis. Frozen samples of 80-100 isolated 15-17 day old fly heads were digested in 10 ml of 5% nitric acid, with a microwave sample digester CEM Mars 2000 using a modified oyster tissue method. Cu was measured using ICP-MS on a Thermo-elemental X5 instrument. Standard solutions were prepared in 5% HNO 3 . Cu65 was measured and each sample was assayed 3 times per ICP-MS run (variation 5-8%). For the wildtype oreR and Ab42 flies, the average of 4 independent pools was calculated, showing an average maximum variation around 25-30%, which includes 15% inherent variation of the ICP-MS method. In cultured cells, growth media, and wash media, copper concentrations were determined by electrothermal heated graphite atomizer (HGA) atomic absorption spectrophotometry with a Perkin Elmer Zeeman Model 5100 instrument. Previously described methods [24] were optimized for the instrument, cells, and media specimens. Counted cell populations of 229610 6 cells were washed with low Cu media and transferred to Eppendorf Safe-Lock (1.5 ml) tubes for digestion. Cells were dried for 1 hour in a heating block at 80uC. The pellet was digested with 100 mL of double-distilled 70% nitric acid (GFS Chemical, Powell, OH), capped and heated overnight at 80uC. The cell digests were diluted with deionized/distilled water prior to analysis. Media was analyzed directly with the addition of an ammonium nitrate modifier to reduce NaCl background interference. A standard reference material that has a known and certified copper concentration (NIST 1571), was purchased from the National Institute for Standards and Technology, and used as quality control specimen. Blanks were also analyzed with each set of media and cell samples. Plasticware used for specimen collection, processing and analysis was acid soaked and rinsed with distilled water prior to use.

Protein Analysis
Cultured cells were lysed in lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 10 mM Na 4 P 2 O 7 , 2 mM Na 3 VO 4 , 1% Triton X-100, 1 mM PMSF, 20 mg/ml aprotinin). For immunoprecipitation (IP), 0.5,1.0 mg of total cellular protein was incubated with primary antibody at 4uC overnight, followed by the addition of Protein A/G-Sepharoses and additional incubation at 4uC for 1 h, then resolved by Tris-Tricine gel or SDS-PAGE. Western blotting of cell samples was performed following SDS-PAGE or Tris-Tricine gel electrophoresis and transfer to polyvinylidene difluoride membranes (Milllipore, Billerica, MA). Blots were incubated with primary antibodies, diluted in 5% milk, after which the blots were washed in Tris-buffered saline containing 0.05% Tween 20 and subsequently incubated with horseradish peroxidase-conjugated secondary antibodies. Immunoblots were developed with an enhanced chemiluminescence kit (Western Lightning; PerkinElmer Life Science, Waltham, MA).

Transient Transfection and Stable Expression
All cell lines were transiently transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). To generate the stable expressing human APP695 in HEK cells, we transfected with pcDNA-APP695, selected by immunoblotting and maintained positive clones in the presence of 200 mg/ml G418 (Gibco-Invitrogen, Carlsbad, CA). To generate the stable expressing human APP695 in FKBP52 knockout mouse embryonic fibroblast (MEF) cells, we cotransfected with pcDNA-APP695 and pBabe vector, selected by immunoblotting and maintained positive clones in the presence of 3 mg/ml puromycin (Invitrogen, Carlsbad, CA).

ELISA for Ab Peptides
Levels of Ab peptides in Drosophila heads shown in Fig. 1H were analyzed as described in [21]. For the analysis of Ab peptides in mammalian cells, stable APP695-HEK cells or stable APP695-FKBP52 knockout MEF cells were transiently transfected with vector or FKBP52-V5 plasmid for 48 hrs. The medium was changed and the conditioned medium was collected. Medium was centrifuged for 10 min at 13000 rpm to remove cellular debris and adjusted to 25 mM Tris-HCl (pH 7.5) containing 1 mM PMSF, 20 mg/ml aprotinin and 0.5% TritonX100. To measure concentration of Ab40 or Ab42, medium was analyzed using an ELISA kit (Covance Research Products Inc. Dedham, MA).

FKBP52 Modifies the Toxic Effects and Levels of Ab Peptides in Drosophila melanogaster
We first investigated the effect of FKBP52 on the toxicity of Ab peptides, by a functional genetic approach. We have previously engineered transgenic Drosophila melanogaster strains that express human Ab42 peptides and shown that Ab expression promotes degeneration of the nervous system with concomitant learning and memory defects, in a dose dependent manner [22,25]. When Ab is expressed using an eye-specific promoter, it induces a rough eye phenotype characteristic of inappropriate organization of ommatidia ( Fig. 1A-B), whereas expression in the CNS causes reduced lifespan (Fig. 1E). These phenotypes are caused by progressive degeneration of the eye and brain tissue, respectively. We performed genetic interaction analysis in these Ab-expressing flies by analyzing the modification of Ab-induced phenotypes in the presence of FKBP52 mutations. Drosophila encodes four known members of the FKBP family [20]. Two of those are homologs of the smaller form (FKBP12, FKBP13), one resembles the atypical form FKBP39 and the fourth member, dFKBP59 [26], is a homolog of FKBP52, containing three FK506 binding domains and three tetratricopeptide repeats (TPR).
Shown in Table 1 are Drosophila strains that carry mutations caused by transposable element insertions in the Drosophila dFKBP59 gene. We generated flies co-expressing Ab42 with each of these dFKBP59 mutations and quantified their Ab42-induced rough eye phenotype (as described in Methods and Fig. 1SA-C). We found that the loss-of-function mutation dFKBP59 c01413 caused enhancement of this phenotype (Fig. 1C), whereas the mutant dFKBP59 EY03538 , which over-expresses dFKBP59, suppressed the Ab42-induced rough eye phenotype (Fig. 1D). Two additional loss of function dFKBP59 mutations were examined, one causing a strong enhancement (dFKBP59 k00424 ) and the second causing a marginal enhancement (dFKBP59 k09010 ) of Ab toxicity (Table 1). All loss of function alleles of dFKBP59 have wild-type eyes (Fig. 1A,  Fig. S1E-F). To further examine this effect, we also tested if dFKBP59 mutations might alter the lifespan of Ab42-expressing flies. Figures 1F-G show a comparison between Ab42-expressing flies with or without a mutation in dFKBP59. In the presence of the gain-of-function mutation dFKBP59 EY03538 , the lifespan of Ab42expressing flies was extended significantly ( Fig. 1F; P = 0.0001), suggesting that the dFKBP59 mutation had a beneficial effect on Ab42 toxicity. We also tested the loss-of-function mutation dFKBP59 c01413 in the same assay and found that it caused the opposite effect ( Fig. 1G; P = 0.04), although this was milder than the effect of the gain of function mutation. This could be due to different expressivity of the mutations.
Our results indicate that FKBP52/dFKBP59 may function independently of its role as a prolyl-isomerase to modify Ab toxicity, as Ab does not contain proline residues. Alternatively, dFKBP59 might indirectly affect Ab phenotypes. In order to further analyze the dFKBP659 effects on Ab42, we measured levels of Ab42 peptides in Drosophila heads over-expressing Ab and dFKBP59. As shown in Fig. 1H, these flies (Abeta/FKBP59gof) have significantly lower steady state levels of Ab peptides, consistent with their suppressed eye and lifespan phenotypes. The results of our genetic analysis suggest that dFKBP59 can modify the toxic phenotypes of Ab42 peptides by affecting the levels of the peptide.

Copper Homeostasis Is Linked to Toxicity of Ab42 in Flies
Our previous studies indicated a role for FKBP52 in the regulation of intracellular copper metabolism [12]. Since altered copper homeostasis has clearly been shown to have a role in Alzheimer's disease [reviewed in 15], we examined whether it might impinge on the toxicity of Ab in our model system. For this, we manipulated copper levels in Drosophila using two independent approaches and examined effects on Ab42-induced phenotypes. In the first approach, we increased or decreased available copper by either raising flies in food supplemented with copper or the metal chelator BCS, respectively. In the second approach, we examined toxicity of Ab in genetic backgrounds where copper transporter genes were mutated.
Copper-supplemented food does not affect the morphology of the eyes of wild type flies ( Fig. 2A). However, Ab42 flies raised on 1 mM copper had more severe rough eye phenotypes (Fig. 2C) than Ab42 flies raised on normal food (Fig. 2B). Fig. 2E shows that ,70% of flies fed 1 mM copper had severe rough eyes, whereas only ,20% of flies raised on normal food had severe rough eyes, suggesting that copper increased the toxicity of Ab42 peptides. In order to reduce the amount of available copper, we raised flies on food supplemented with the chelator BCS. Since BCS shows minimal toxicity when fed to flies (data not shown), we fed 1 mM and 5 mM BCS to freshly eclosed flies with eye-specific Ab42 expression. We examined flies aged to 25 days old and found that BCS feeding ameliorated the Ab-induced rough eye phenotype (Fig. 2D). Control flies raised on normal food had only moderate and severe eye phenotypes whereas flies fed 1 mM or 5 mM BCS also had mild phenotypes (18% and 25% of the progeny, respectively; Fig. 2F). These experiments demonstrate that altered Cu levels can directly alter Ab toxicity in flies.
In order to quantify the increase of Cu levels in flies raised on copper-supplemented food, we used ICP-MS to measure copper in fly heads. As shown in Fig. 2G, supplementing the fly food with 0.03 mM or 0.3 mM of copper causes an increase in the steady state levels of copper in Drosophila heads, in a dose dependent manner (6,340 ng/g in 0.03CumM and 13,860 ng/g in 0.3 mM copper, as compared to 3,500 ng/g in flies fed normal food). Expression of Ab42 does not affect the steady state levels of copper in fly heads, which contain less than one third of copper compared to the rest of the body (Fig. S2B). More than 75% of the metal is found in the insoluble fraction of head extracts (Fig. S2C).

Mutations in the Copper Transporters Atox1 and Ctr1A Modify Ab42-Induced Rough Eye Phenotypes and Alter Levels of Copper
In order to further analyze the interaction of copper with Ab, we tested the effects of the Cu transporter genes Ctr1A and Atox1 on Ab phenotypes. Flies express three isoforms of Copper transporter 1 [Ctr1; 27; reviewed in 28], which is a transmembrane protein responsible for import of copper ions into the cell [reviewed in 29]. In addition, Drosophila carries a homolog of Atox1 [20], which is a cytoplasmic chaperone responsible for delivery of copper to copper transporting ATPases and hence to the secretory pathway.
We first examined the rough-eye phenotype of Drosophila overexpressing Ctr1A and Ab42 and found that it was enhanced, compared to flies expressing only Ab42 (Table 1). Given the role of Ctr1A in importing Cu into the cell, we hypothesized that the Ctr1A effects might be mediated by higher levels of available copper. To test this, we measured the amount of copper in heads of flies overexpressing the copper transporters Ctr1A under the control of the UAS/Gal4 system [30] and found a 4-fold increase in the levels of copper, as compared to control heads (Fig. 3A). The fact that overexpression of Ctr1A caused elevated levels of copper supports our hypothesis that the enhancement of the Ab42 phenotype when this transporter is over-expressed may be due to the presence of elevated copper levels in the cell.
We then examined how mutations in Atox1 are affecting the Ab42 phenotype. Table 1 shows that three different loss-offunction alleles of the Atox1 gene, Atox1 e01272 , Atox1 EY15780 , and Atox1 f00729 cause enhancement of the Ab42 rough eye phenotype. Atox1 loss of function mutations do not affect eye morphology (Fig. S1G-I). Since Atox1 is involved in the delivery of copper to copper-transporting ATPases in the trans Golgi network (TGN), perturbations in its expression would disturb the proper localization of copper in the cell, potentially exacerbating Ab toxicity.
Supporting this, we found that Atox1 loss-of-function mutants had more copper than control flies (Fig. 3B). Since we have previously shown that dFKBP59 interacts with Atox1 and is involved in copper efflux [12], we tested levels of copper in loss of function dFKBP59 mutant Drosophila heads and found that they had increased levels of copper (Fig. 3B). Although all loss of function mutations of Atox1 and dFKBP59 that we examined increased levels of copper, the increases were of varying degrees, perhaps reflecting genetic background effects. It is worth noting however, that the three mutants with the strongest phenotypic effects on Ab42, Atox1 EY15780 , Atox1 f00729 and dFKBP59 c01413 , showed the higher increases in levels of copper.
We showed the same effects of loss of FKBP52 function on copper levels, in mammalian cells. Changes in copper levels can be measured in MEF cells treated with 50 mM or 100 mM of the chelator BCS, as shown in Fig. 3C. Using this assay, we showed that immortalized MEF cells obtained from FKBP52 null mice [FKBP52(2/2)] had elevated copper compared to wild type MEF cells (48 pmol/10 6 cells versus 31 pmol/10 6 cells; p = 0.05; Fig. 3D), confirming our previous observations that overexpression of FKBP52 causes lower levels of copper [12]. These experiments further support the involvement of FKBP52 in copper homeostasis and suggest that changes in levels of cellular copper by mutations in FKBP52 may potentiate the toxic effects of Ab.

Synergy between Immunophilin FKBP52 and Copper Homeostasis Moderates Ab42 Toxicity
We next performed a genetic interaction analysis, to test whether the effect of dFKBP59 on Ab toxicity might be dependent on copper homeostasis. If this were true, we should be able to detect an epistatic relationship between available copper levels and functions of dFKBP59, using as a read-out the Ab42-induced phenotypes in flies. We chose to manipulate the levels of copper by raising Drosophila on BCS-supplemented food, a treatment that, as we showed above, improves Ab42 phenotypes. Our experimental progeny was expressing Ab42 in the presence of either overexpression or loss-of-function mutations of dFKBP59 and was tested for effects on the lifespan phenotype, which provides a measure of CNS function.
As a control, we treated Ab-expressing flies with BCS and observed a beneficial effect on their lifespan (Fig. 4A). We next treated Ab flies carrying a loss-of-function mutation of dFKBP59, which showed a highly significant increase in lifespan, relative to non-BCS treated flies ( Fig. 4B; P = 0.0001). Similarly, treated Ab flies over-expressing dFKBP59 showed significantly increased lifespan versus non-treated flies ( Fig. 4C; P = 0.0004). The fact that this increase in lifespan is more evident towards the later stages of life may indicate that the effects are dependent on specific levels and/or nature of Ab oligomeric species. These results suggest that levels of intracellular copper influence the effects of dFKBP59 on Ab42 toxicity.

Role of FKBP52 in Ab Metabolism in Mammalian Cells
We next investigated whether FKBP52 affected Ab levels in mammalian cells expressing APP. Previous studies have suggested a role for other members of the prolyl-isomerase family, such as FKBP12 and Pin1, in promoting the non-amyloidogenic processing of APP [reviewed in 31]. However, no previous study has addressed the involvement of the larger immunophilin FKBP52 in this pathway.
To test the effect of FKBP52, we transiently transfected FKBP52 in HEK cells that had stable expression of APP695 (Fig. 4D) and assayed levels of Ab in the conditioned media with an ELISA assay. Fig. 4E shows that overexpression of FKBP52 reduced the levels of Ab42 peptides in stable APP695-HEK cells. We also transfected human FKBP52 in FKBP52-knockout APP695-MEF cells (Fig. 4F) and found that in these cells, Ab42 levels were also less than those in FKBP52 knockout cells not transfected with human FKBP52 (Fig. 4G). The fact that HEK cells with wild-type FKBP52 and MEF cells with FKBP52 knockout produce the same amount of Ab may be due to different steady state levels of Ab in HEK versus MEF cells. These results suggest that, as in Drosophila, FKBP52 reduces levels of Ab in mammalian cells as well.

FKBP52 Interacts with APP
We subsequently investigated whether FKBP52 and APP interact physically. We co-transfected Myc-APP695-FLAG and FKBP52-V5 into HEK cells. As is seen in Fig. 5A, using immunoprecipitation with appropriate antibodies, we were able to detect an interaction between the two proteins. We confirmed this interaction by also testing the pools of endogenous proteins in these cells. As HEK cells express both endogenous APP and FKBP52, they are suitable for testing the in vivo interaction of APP and FKBP52 (Fig. 5B). Using co-immunoprecipitation, we detected evidence for an interaction between endogenous APP and FKBP52, in cell lysates of HEK cells (Fig. 5C), confirming the results that we obtained with the transfected forms of the two proteins. This interaction was blocked after treatment of the cells with 0.5 mM FK506 (Fig. 5C). Thus, similar to the smaller immunophilin, FKBP12, the larger protein FKBP52 is interacting with APP.
We next sought to determine which domain of FKBP52 is responsible for the interaction with APP. Guided by the interaction of the PPIase domain of FKBP12 with APP, we cloned and transfected the truncated domain I-II of FKBP52, which encompasses amino-acids 1-261 and contains a PPIase domain. As shown in Fig. 5D, the truncated domain of FKBP52 bound APP in a similar fashion to the binding of the full length protein, suggesting that the interaction of FKBP52 with APP is mediated by the PPIase domain. Domain I of FKBP52 associates with various cellular factors, including the glucocorticoid receptor [1,32] dynein [33], the transient receptor potential channel [TRP; 34], and Atox1 [12]. With the exception of dynein, most of these factors dissociate from FKBP52 after addition of FK506. Since we showed that FK506 can promote dissociation of the endogenous APP/FKBP52 complex (Fig. 5C), we suggest that the PPIase domain plays a direct role in the FKBP52/APP interaction.

Discussion
The high molecular weight immunophilin FKBP52 belongs to a family of versatile multi-domain proteins that are abundantly expressed in the nervous system and often show increased expression in damaged or degenerating brain regions. In the present study, we have identified a novel function of FKBP52 in Ab-mediated toxicity using a genetic model in Drosophila that expresses Ab42 peptides. The protective effect of FKBP52 on Ab toxicity during Drosophila aging was evident from the observations that FKBP52 loss of function mutations potentiated Ab toxicity, while over-expression of FKBP52 delayed or suppressed Abinduced phenotypes. Moreover, through genetic interactions and chemical approaches, we also found evidence that the effects of FKBP52 may be modulated by changes in intracellular copper homeostasis. These observations are consistent with our previous reports that FKBP52 interacts with the copper efflux machinery, and as such, establish a new aspect of involvement of the immunophilin family in Alzheimer's-related mechanisms. Several lines of evidence link the protective effects of FKBP52 with intracellular copper homeostasis. First, FKBP52 directly interacts with the copper metallochaperone Atox1 [12], a protein that delivers copper to the copper transporting ATPAses ATP7A and ATP7B [reviewed in 35]. Second, through genetic screens in this study, we found that mutations in the copper transport genes Ctr1A and Atox1, which directly regulate intracellular copper levels, modify Ab-induced phenotypes in Drosophila. Third, chemical manipulation of dietary copper levels also decreases or increases, respectively, the protective effect of FKBP52 on Ab toxicity. Finally, MEF cells isolated from FKBP52(2/2) mice show increased levels of copper compared to wild type MEF cells and over-expression of FKBP52 causes efflux of copper [12].
The metallobiology of copper plays a significant role in several neurodegenerative conditions. Interestingly, copper influences the aggregation properties of ''toxic peptides'' that contribute to these conditions, including beta amyloid, prion protein and a-synuclein as all three of these agents can bind copper ions [reviewed in 36]. The presence of copper in mildly acidic conditions induces aggregation of the Ab peptide [13] and may exacerbate pathology linked to Ab deposition. Our experiments show that all genetic manipulations that increased levels of copper, also enhanced the Ab phenotypes. Copper manipulations were mediated either by increased copper entry, through over-expression of the plasma membrane transporter Ctr1A, or by misregulation of cytoplasmic copper trafficking through loss-of-function mutations of the Atox1 and FKBP52 genes. The enhancement of Ab phenotypes by increasing dietary copper also supports these observations. Since copper is delivered to the trans-Golgi network (TGN) by the cytoplasmic transporter Atox1, we hypothesize that the interaction of Ab with copper may at least partially take place in the TGN. Given the mildly acidic pH of the TGN [37][38], increased levels of copper in this compartment would lead to enhancement of Ab toxicity and result in more severe Ab-induced phenotypes, possibly through the induction of oxidative stress. Supporting the role of oxidative damage, over-expression of the anti-oxidative stress gene ferritin heavy chain suppressed the Ab42-induced short lifespan in Drosophila [39]. We further found that flies over-expressing dFKBP59 had lower levels of Ab peptides, consistent with their suppressed phenotypes. Based on this, we suggest that dFKBP59 over-expression leads to increased Ab turnover.
The interaction of FKBP52 with the transporter Atox1 presents a novel aspect of copper metabolism. FKBP52 participates in many cellular processes, including the translocation of steroid receptor complexes to the nucleus through interactions with dynein [reviewed in 40]. FKBP52 also has chaperone activity shown by suppression of the aggregation of heat-denatured citrate synthase [41]. We propose that FKBP52 may be required for the proper function of Atox1. Further analysis using double mutants of Atox1 and FKBP52 and examining their effects on Ab toxicity would be needed in order to confirm this hypothesis.
We also examined the effects of FKBP52 in mammalian cells expressing APP. Unlike Ab, which does not contain proline amino acids and is not regulated by prolyl isomerization, the APP holoenzyme binds the prolyl-isomerase Pin1 in its intracellular tail [10] and it also interacts with the small immunophilin FKBP12 [11]. In the current studies we provide evidence that APP also binds FKBP52, via its FK506 binding domain and that FKBP52(2/2) cells have higher levels of Ab peptides than the same knock-out cells reconstituted with FKBP52. The physical interaction of FKBP52 with APP suggests that this large immunophilin, in addition to altering Ab levels, may have a role on the metabolism of APP. The effects of smaller members of the immunophilin family on the processing of APP would support such a role and future experimentation will address this hypothesis. In support of this novel role of immunophilins, we showed recently that a mutation in the Drosophila Ryanodine receptor homolog Rya-r44F could modify a APP-overexpression associated phenotype [42]. The FKBP12 protein interacts with ryanodine receptors [43], further implicating signaling through the immunophilin famlily with APP metabolism.
In summary, our studies show that the large immunophilin FKBP52 modulates Ab toxicity, possibly through a mechanism that involves homeostasis of cellular copper. Our data does not rule out the possibility that the effects of FKBP52 mutations and metal transport act in parallel pathways, however, it provides indirect evidence for a possible mechanistic link between these respective pathways. Examination of effects of Ab in FKBP52 knock-out mice will further validate our observations. The function of immunophilins is modulated by the FK506 family of ligands, several members of which have been developed to bind their targets without causing immune suppression. It will be interesting to examine if such ligands can modify the interaction of FKBP52 with APP and Ab. We have evidence that the FK506binding domain of FKBP52 is involved in the binding with APP, suggesting that immunophilin ligands may interfere with this interaction. Further studies will show whether these ligands are also involved in modulating toxicity of Ab and may open the field for the development of a novel class of agents against Alzheimer's disease. RNA levels in fly heads over-expressing FKBP59 EY03538 compared to control flies, measured by real-time PCR analysis, as described in [21]. (B) Cu concentration in pools of 100 flies, aged to 15 days.

Supporting Information
Bodies contain the majority of Cu. (C) Cu concentration in pools of 60-80 flies, aged to 15 days. Control and flies expressing one or two copies of the Ab42 transgene were analyzed. From each sample, the supernatant (containing the soluble fraction) and pellet (containing the insoluble fraction) were analyzed separately. The majority of Cu is found in the insoluble fraction. Found at: doi:10.1371/journal.pone.0008626.s002 (0.79 MB TIF)