The Fanconi Anemia Proteins Functionally Interact with the Protein Kinase Regulated by RNA (PKR)*

Protein kinase regulated by RNA (PKR) plays critical roles in cell growth and apoptosis and is implicated as a potential pathogenic factor of Alzheimer's, Parkinson's, and Huntington's diseases. Here we report that this proapoptotic kinase is also involved in Fanconi anemia (FA), a disease characterized by bone marrow (BM) failure and leukemia. We have used a BM extract to show that three FA proteins, FANCA, FANCC, and FANCG, functionally interact with the PKR kinase, which in turn regulates translational control. By using a combined immunoprecipitation and reconstituted kinase assay, in which an active PKR kinase complex was captured from a normal cell extract, we demonstrated functional interactions between the FA proteins and the PKR kinase. In primary human BM cells, mutations in the FANCA, FANCC, and FANCG genes markedly increase the amount of PKR bound to FANCC, and this PKR accumulation is correlated with elevated PKR activation and hypersensitivity of BM progenitor cells to growth repression mediated by the inhibitory cytokines interferon-γ and tumor necrosis factor-α. Specific inhibition of PKR by 2-aminopurine in these FA BM cells attenuates PKR activation and apoptosis induction. In lymphoblasts derived from an FA-C patient, overexpression of a dominant negative mutant PKR (PKRK296R) suppressed PKR activation and apoptosis induced by interferon-γ and tumor necrosis factor-α. Furthermore, by using genetically matched wild-type and PKR-null cells, we demonstrated that forced expression of a patient-derived FA-C mutant (FANCCL554P) augmented double-stranded RNA-induced PKR activation and cell death. Thus, inappropriate activation of PKR as a consequence of certain FA mutations might play a role in bone marrow failure that frequently occurred in FA.

The Protein kinase regulated by RNA (PKR) 1 plays a critical role in translational control (1). PKR has been subjected to intensive investigations (2) because of its antiviral function. The underlying mechanism by which PKR mediates antiviral defense in mammalian cells is the inhibition of translation through phosphorylation of eukaryotic initiation factor 2 ␣-subunit (eIF-2␣) (2). It is known that PKR becomes autophosphorylated on multiple serine/threonine residues following viral infection by dsRNA produced by viral gene expression or viral replication. Upon activation by autophosphorylation, PKR phosphorylates Ser-51 on eIF-2␣, which in turn blocks the initiation of protein synthesis (3). Another important part of the cellular functions of PKR is its role in regulation of cell proliferation, survival, and apoptosis. Because of its pro-apoptotic effects, forced expression of PKR in mouse, insect, and yeast cells causes translational inhibition and cell death (4,5). Indeed, inappropriate activation of PKR has been associated with certain disease states characterized by high levels of apoptosis. For example, PKR and the PKR-like kinase PERK have been implicated as important pathogenic factors in the Alzheimer's, Parkinson's, and Huntington's diseases (6 -9).
Research on Fanconi anemia (FA) has recently generated great interest because the disease serves as an excellent model for hematopoietic failure and leukemic evolution, and because FA proteins function in cellular responses to a variety of stresses including signals of DNA damage and apoptosis. FA is an autosomal recessive disease characterized by progressive bone marrow failure, variable congenital anomalies, and a predisposition to cancer (10 -12). Cells from FA patients are hypersensitive to DNA cross-linking agents such as mitomycin C and diepoxybutane. FA is genetically heterogeneous, with at least 11 complementation groups identified thus far (13). The genes encoding the groups A (FANCA), C (FANCC), G (FANCG), D1 (FANCD1/BRCA2), D2 (FANCD2), E (FANCE), F (FANCF), and L (FANCL) have been cloned (14 -21). The biological functions of these FA gene products remain mostly unknown.
We demonstrated previously that one of the FA proteins, FANCC, functions to protect cells from cytotoxicity mediated by PKR (22,23). Because PKR was found to be constitutively activated in FA cells with mutations in the FANCC gene (22), we reasoned that activation of PKR in bone marrow may play a role in the pathogenesis of bone marrow failure in children with FA. We thus sought to investigate the mechanism by which FANCC modulates PKR activity. We found that suppression of the PKR kinase in FA cells required a coordinated action of FANCC and the molecular chaperone Hsp70, a cytoprotective factor we have previously shown to act in concert with FANCC to protect hematopoietic cells from cytotoxicity induced by mitogenic inhibitors interferon ␥ (IFN-␥) and tumor necrosis factor ␣ (TNF-␣) (24). In this report, we demonstrate that three FA proteins, FANCA, FANCC, and FANCG, functionally interact with the PKR kinase. Mutations in these FA genes cause abnormal accumulation of PKR kinase bound to the FANCC protein, which is correlated with elevated PKR activation and hypersensitivity of BM progenitor cells to growth repression mediated by inhibitory cytokines IFN-␥ and TNF-␣. Our results suggest that inappropriate activation of PKR as a consequence of certain FA mutations might play a role in bone marrow failure frequently occurring in FA.

EXPERIMENTAL PROCEDURES
Cell Culture, Patient Material, and Treatments-Normal and mutant lymphoblast lines derived from patients diagnosed with Fanconi anemia (FA) were maintained in RPMI media 1640 (Invitrogen) supplemented with 15% fetal bovine serum. Cells were stimulated with recombinant human IFN-␥ (10 ng/ml) and TNF-␣ (10 ng/ml) (R & D Systems) for 16 h. Mouse embryo fibroblasts (MEF) from wild-type (PKR ϩ/ϩ ) and PKR knockout (PKR o/o ) mice (25), were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum. MEF were transfected with 50 g/ml of poly(rI⅐rC) (dsRNA) (Amersham Biosciences) and lysed 24 h post-transfection. Bone marrow mononuclear cells (BM-MNCs) of healthy donors and patients with a diagnosis of FA in complementation groups A or G were obtained in accordance to guidelines from the Institutional Review Board of the Cincinnati Children's Hospital Medical Center. BM-MNCs were cultured in RPMI 1640 medium containing 20% dialyzed fetal bovine serum and supplemented with stem cell factor (100 ng/ml), thrombopoietin (100 ng/ml), and G-CSF (100 ng/ml). All growth factors were purchased from R & D Systems. On day 3 of culture, cells were stimulated with IFN-␥ and TNF-␣ (1 ng/ml each) for 16 h before lysis.  27) was constructed by site-directed mutagenesis by using SF␤91-PKR as the template and primers (sense, 5Ј-GGAAAGACTTACGTT-ATTAGACGTGTTAAATATAATAAC-3Ј; antisense, 5Ј-GTTATTATA-TTTAACACGTCTAATAACGTAAGTCTTTCC-3Ј). The FA-C patientderived mutant FANCC L554P cDNA (28) was removed from pLXSN-L554P (24) and subcloned into the NotI site of SF␤91 as described above. The SF␤91 plasmids (10 g each) were used to produce retroviral supernatant. Lymphoblasts or MEFs were exposed to the retroviral supernatants in the presence of 8 g/ml Polybrene (Sigma).

Construction of Retroviral Expression Vectors and Transduction-
Immunoprecipitation and Immunoblotting-Whole cell extracts (WCE), prepared in Nonidet P-40 lysis buffer (24), were precleared with 50 l of 50% protein A-Sepharose suspension (Amersham Biosciences) for 1 h at 4°C. After separation of the protein A-Sepharose, the extract was incubated with the indicated primary antibodies. Immunocomplexes were then bound to protein A-Sepharose beads, recovered by centrifugation, and washed with Nonidet P-40 lysis buffer. For immunoprecipitation with BM-MNC extracts, the FANCC antibodies were conjugated to the M-280 paramagnetic Dynabeads (sheep anti-rabbit IgG; Dynal) and incubated with 300 -500 g of BM-MNC extracts (at a ratio of about 100 g of extract proteins to 10 l of anti-FANCC beads) for 2 h at 4°C. Beads were washed three times and collected on a magnetic particle concentrator (Dynal). The supernatant, first wash, and the beads were subjected to immunoblot analysis. For immunoblotting, samples were separated by SDS-PAGE and transferred onto nitrocellulose membranes. Immunoblots were incubated with the indicated primary antibodies for 4 h at room temperature or overnight at 4°C. After washing, the blots were incubated with appropriate secondary antibodies for 2-4 h at room temperature and developed using an enhanced chemiluminescence kit (Amersham Biosciences).
Reconstituted in Vitro Kinase Assays-Reconstituted PKR kinase reactions were carried out in kinase buffer containing 20 mM Tris-HCl, pH 7.5, 50 mM KCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 1 mM ATP, 5% glycerol, and protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, and 1 g/ml leupeptin). Unless indicated otherwise, each reconstituted reaction mixture contained 3 l of the anti-FANCC beads, 2 l of each purified protein, and Hsp70 (5 M, StressGen). The mixtures were incubated with 50 ng of eIF2␣ (gift from Dr. B. Batta, University of Nebraska) at 30°C for 10 min. Then 5 g/ml of poly(rI⅐rC) (Amersham Biosciences) were added, and the mixtures were incubated for an additional 10 min. The levels of PKR and eIF-2␣ phosphorylation were analyzed by using the respective phospho-specific PKR (BioSource International Inc., Camarillo, CA) and eIF-2␣ (Cell Signaling Technology, Beverly, MA) antibodies.
In Vivo 32 P and 35 S Labeling of Proteins-Cells were starved for 60 min in phosphate-free medium containing 10% dialyzed fetal bovine serum and were treated with a combination of IFN-␥ and TNF-␣ (10 ng/ml each, R & D Systems) for 2 h. Labeling was performed in the same medium by addition of [ 32 P]orthophosphate (150 Ci/ml, PerkinElmer Life Sciences). After labeling for 3 h, whole cell lysates were prepared as described above and were subjected to immunoprecipitation with a monoclonal antibody specific for human PKR conjugated to agarose. [ 32 P]Phosphate-incorporated immunocomplexes were analyzed by SDS-PAGE followed by autoradiography. Immunoblot analysis was also performed on these samples to determine the quantity of PKR precipitated by the anti-PKR antibody. For 35 S labeling, 5 ϫ 10 6 lymphoblasts were cultured in the presence of IFN-␥ and TNF-␣ (10 ng/ml each, R & D Systems) for 15 h. Cells were rinsed with methionine-cysteine-free medium, and labeling was again performed in the same medium containing 50 Ci/ml [ 35 S]methioninecysteine labeling mix (PerkinElmer Life Sciences) and IFN-␥ and TNF-␣ (10 ng/ml each, R & D Systems) and incubated at 37°C for an additional hour. Whole cell lysates were prepared in Nonidet P-40 lysis buffer, and protein concentration was determined. Protein synthesis was measured by the incorporation of [ 35  CFU-Es and BFU-Es were cultured in either the absence or presence of IFN-␥ and TNF-␣ (1 ng/ml each, R & D Systems) at 37°C in 5% CO 2 in air and counted after 7 and 14 days, respectively, using a dissecting microscope. Colony growth results were expressed as mean (of triplicate plates) Ϯ S.D. colonies per 10 5 cells.
Analysis of Cell Viability and Apoptosis-Cell viability was measured by trypan blue exclusion analysis. To quantify apoptotic cells, we used a polyclonal antibody to the active form of caspase 3 in a flow cytometric assay to detect cells in the early stages of apoptosis. Caspase 3 is activated during apoptosis to varying degrees in different cell types, and recent studies (29) indicate that the apoptotic response of FA mutant cells is caspase 8-and 3-dependent. The active, cleaved form of caspase 3 therefore provides an instructive and biologically relevant marker of FA mutant cells undergoing programmed cell death. Cells were harvested, washed with phosphate-buffered saline, and resuspended at a concentration of 1 ϫ 10 6 /ml in staining buffer containing phosphate-buffered saline, 2% fetal calf serum, 0.1% sodium azide. This cell suspension (400 l) was then added to 400 l of Cytofix/Cytoperm (Pharmingen) and incubated for 20 min on ice to fix and permeabilize the cells. The cells were then washed with 2 ml of Perm/Wash buffer (Pharmingen) and resuspended in 100 l of Perm/Wash buffer. Purified rabbit immunoglobulin G (IgG; 10 l; Pharmingen) was then added to each sample as a blocking antibody to prevent the nonspecific uptake of fluorochrome-conjugated antibody. After 15 min, 20 l of phycoerythrin-conjugated anti-active caspase 3 antibody (Pharmingen) was added, and samples were incubated for 30 min in the dark at room temperature. Cells were washed with 2 ml of Perm/Wash buffer, resuspended in 500 l of staining buffer, and analyzed by flow cytometry by using a FACSCalibur (BD Biosciences). Camptothecin-treated cells were used as a positive control.

Mutations in FANCA, FANCC, and FANCG Genes Lead to Abnormal Activation of PKR and Translational Repression-
The increased association of the PKR kinase with the FANCC protein in lymphoblasts derived from an FA patient carrying an FANCC mutation (23) 1B). This would be expected because the FANCC protein in these FA cells is presumably normal and capable of binding to Hsp70 (24).
Hematopoietic cells from FA patients and Fancc knockout mice are hypersensitive to the treatment of the mitogenic inhibitors IFN-␥ and TNF-␣ (11, 22, 29 -31). We and others (22,32) have demonstrated that these inhibitory cytokines can induce apoptosis in FA cells and other cells through the activation of the pro-apoptotic protein kinase PKR and subsequent inhibition of protein synthesis. We sought to determine whether stimulation with IFN-␥ and TNF-␣ could overactivate the PKR kinase in lymphoblasts derived from FA patients carrying FANCA, FANCC, or FANCG mutations. Fig. 2A shows the input controls of whole cell extracts of normal and FA lymphoblasts used in these experiments. Significantly, treatment of the FANCAϪ/Ϫ, FANCCϪ/Ϫ, and FANCGϪ/Ϫ lymphoblasts with IFN-␥ and TNF-␣ further increased the accumulation of PKR on FANCC in FANCCϪ/Ϫ, FANCAϪ/Ϫ, and FANCGϪ/Ϫ lymphoblasts (Fig. 2B, lanes 4, 6, and 8). To determine whether the increased amounts of PKR bound to FANCC might be associated with PKR activation, we labeled the lymphoblasts with [ 32 P]orthophosphate and measured levels of 32 P-labeled PKR immunoprecipitated with a monoclonal PKR antibody. As expected, treatment of normal and FA lymphoblasts with IFN-␥ and TNF-␣ induced PKR phosphorylation (Fig. 2C). However, IFN-␥ and TNF-␣ stimulation led to a much higher PKR activation in FANCAϪ/Ϫ, FANCCϪ/Ϫ, and  (Fig. 2B). Therefore, these results provide direct evidence linking the nonproductive interaction between PKR and the FA proteins to deregulation of the proapoptotic protein kinase PKR.
Because PKR activation leads to translational repression (1, 2), we reasoned that global translational activity in FANCAϪ/Ϫ, FANCCϪ/Ϫ, and FANCGϪ/Ϫ cells should be reduced to a greater extent than in normal cells after exposure to IFN-␥ and TNF-␣. Protein synthesis was significantly reduced by IFN-␥ and TNF-␣ treatment in FANCAϪ/Ϫ, FANCCϪ/Ϫ, and FANCGϪ/Ϫ lymphoblasts, demonstrating a 60 -80% decrease after 16 h (Fig. 2D). Taken together, these results thus support the notion of links between abnormal PKR activation and deregulation of translational control in FA hematopoietic cells well documented to undergo excessive apoptosis in response to IFN-␥ and TNF-␣ (11, 22, 29 -31).
PKR Activity Is Elevated in FA Bone Marrow Cells-To verify the role of PKR signaling in primary human hematopoietic cells, we sought to characterize the interaction between FA proteins and the PKR kinase in BM cells from healthy donors and FA patients. Since the bone marrow uniformly fails in FA patients and abnormalities of other organ systems are sporadic (10, 11), we first chose a normal human BM cell extract that can be assayed for PKR kinase activity but has no overexpressed proteins or interfering elements. To improve the immunoprecipitation, we coated the FANCC antibody with Dynabeads before incubation with the BM cell extracts. As shown in Fig. 3A, essentially all the FANCC protein present in the normal BM mononuclear cells could be pulled down by the antibody beads, as indicated by the absence of FANCC in the supernatant or the wash fraction (Fig. 3A, lanes 4 -6). With rabbit whole IgG as the precipitation control, FANCC was detected in the supernatant but not in the immunoprecipitated fraction (Fig. 3A, lanes 7 and 8). Further analysis of the anti-FANCC immunoprecipitate revealed that detectable amounts of FANCA, FANCG, PKR, and Hsp70 could be found in the fraction bound to the anti-FANCC beads (Fig. 3A, lane 4). To examine PKR-FA protein interaction in primary FA BM cells, two FA patients, genotyped in the complementation groups A (FA-A) and G (FA-G), respectively, were recruited for the study. The results with the immunoprecipitates from the FA-A and FA-G patients were similar to those obtained with the FA mutant cell lines (Fig. 2). Specifically, these immunocomplexes contain significantly more FANCC-bound PKR than that in the immunoprecipitate of the normal control under the same conditions (Fig. 3B). In these FA BM cells, about 25-30% of total PKR was accumulated on the FANCC protein as compared with less than 5% in the normal BM cells (Fig. 3B, compare  lane 4 with lanes 6 and 8). FANCC from both FA-A and FA-G BM-MNCs coimmunoprecipitated with small amounts of Hsp70, suggesting that the FANCC protein in both the FA-A and FA-G BM cells retains the capacity to bind this chaperone protein. However, no coimmunoprecipitated FANCA or FANCG was detected (Fig. 3B, lanes 6 and 8), consistent with reports by others that mutations in the FANCA or FANCG genes abolish the interactions between the FA proteins (12,18).
To examine PKR activation in primary FA BM cells, we determined the levels of PKR and eIF-2␣ phosphorylation in the cell extracts from normal and FA BM cells stimulated with IFN-␥ and TNF-␣, using the respective phospho-specific PKR and eIF-2␣ antibodies. Normal BM cells showed a low level of activated PKR after treatment with IFN-␥ and TNF-␣ (Fig. 3C,  lane 1). However, there was ϳ3-4-fold more phosphorylated PKR detected in the BM cells from the FA patients than in normal BM cells (Fig. 3C, compare lane 1 with lanes 2 and 3). Significantly, the degree to which increased PKR was phosphorylated correlated positively with the accumulation of PKR on the FANCC protein in these BM cells (compare Fig. 3, B with C).
Treatment of FA cells with combinations of either IFN-␥ and double-stranded RNA or IFN-␥ and TNF-␣ dramatically increases phosphorylation of the eukaryotic translation initiation factor eIF-2␣ as a result of PKR activation (22,23). In MCF-7 breast cancer cells, TNF-␣ stimulation causes a PKR-dependent elevation of phosphorylated eIF-2␣ and subsequent translational repression (20). We thus determined the levels of eIF-2␣ phosphorylation in BM cell extracts with a phosphospecific antibody to phosphoserine 51 of eIF-2␣. We observed 2-3fold higher phosphorylated eIF-2␣ in FA BM cell extracts than in normal BM cells (Fig. 3D, top panel, compare lane 1 with  lanes 2 and 3). The amounts of total eIF-2␣ proteins were comparable among these different BM cell extracts (Fig. 3D,  bottom panel).
Apoptosis Induction by IFN-␥ and TNF-␣ or dsRNA in FA Cells Is Dependent on PKR Activation-We next examined the biological consequences of the abnormal PKR activation observed in the FA BM cells. It is known that BM cells from FA patients overproduce IFN-␥ and TNF-␣, which repress the growth of BM progenitor cells through mechanisms involving apoptosis (30,31). To investigate whether de-regulated PKR activation could sensitize the FA BM cells to the inhibitory cytokines, we grew BM-committed progenitor cells in a semisolid assay for the growth of erythroid colony-forming and burst-forming units (CFU-E and BFU-E, respectively) in the presence of IFN-␥ and TNF-␣. In the absence of IFN-␥ and TNF-␣, the normal control had a mean number of CFU-E and BFU-E (1ϫ10 5 seeded BM-MNCs) of 108.4 and 86.6, respectively (Fig. 4, A and B). BM cells from the FA-A patient showed significant inhibition of colony growth with average numbers of 33.2 for CFU-E and 28.8 for BFU-E. When cells were cultured in the presence of IFN-␥ and TNF-␣ at 1 ng/ml each, a dose that does not have an effect on the growth of the normal progenitors, the mean numbers of CFU-E and BFU-E formed by the FA BM cells were significantly decreased to 8.6 and 4.9, respectively (Fig. 4, A and B). Collectively, these results suggest that FA mutations could cause abnormal PKR activation, leading to growth inhibition of hematopoietic progenitors.
To determine specifically the role of PKR activation in eIF-2␣ phosphorylation and apoptotic death in FA BM cells, we treated normal and FA BM cells with IFN-␥ and TNF-␣ in the presence of 2-aminopurine (2-AP), the nucleoside analog known to inhibit the kinase activity and PKR-dependent phosphorylation of eIF-2␣ (33). As shown in Fig. 4C, 2-AP treatments suppressed PKR activation as manifested by the inhibition of both PKR and eIF-2␣ phosphorylation (Fig. 4C, lanes 4 -6). Thus, the ability of 2-AP to block eIF-2␣ phosphorylation indicates that enhanced eIF-2␣ phosphorylation observed in FA BM cells indeed is a downstream effect of PKR activation. We next evaluated the effect of 2-AP on apoptotic induction by IFN-␥ and TNF-␣ in FA-A BM cells. A high concentration of 2-AP (10 mM) alone did not induce significant apoptosis in both normal and FA BM cells (Fig. 4D). However, in cells treated with IFN-␥ and TNF-␣, 2-AP blocked apoptosis induction in FA-A BM cells (Fig. 4D). Thus, the inhibition of apoptosis by 2-AP appeared to be the consequence of its suppression of PKR activation.
To confirm the contribution of PKR activation to the FA

FIG. 3. The activity of PKR kinase in normal and FA bone marrow cells.
A, coimmunoprecipitation of FANCA, FANCC, FANCG, PKR, and Hsp70 from normal BM-MNC extracts. Anti-FANCC antibodies (Ab) were coated onto paramagnetic beads and incubated with 500 g of normal BM-MNC extracts. After washes, the anti-FANCC beads (B), supernatant (S), and first wash (W) were subjected to immunoblot analysis using antibodies against the indicated proteins. apoptotic pathway, we overexpressed a dominant negative mutant, PKR K296R (catalytically inactive; see Ref. 27), in lymphoblasts derived from an FA-C patient and subsequently quantified IFN-␥ and TNF-␣-induced apoptosis. As shown in Fig. 5A, expression of PKR K296R markedly reduced phosphorylation of eIF-2␣ in both normal and FA-C mutant lymphoblasts exposed to IFN-␥ and TNF-␣ (Fig. 5A, lanes 6 and 8). There was a substantial amount of phosphorylated eIF-2␣ in untreated FA-C mutant lymphoblasts (Fig. 5A, lane 3). This likely resulted from the constitutive PKR activation in the FA-C lymphoblasts, as reported previously (23). PKR K296R expression completely abolished this constitutive eIF-2␣ phosphorylation in FA-C mutant lymphoblasts (Fig. 5A, lane 7). Flow cytometric analysis of apoptotic death (Fig. 5B) showed that ϳ40% of empty vector-transduced FA-C cells underwent apoptosis after exposure to IFN-␥ and TNF-␣ (compared with less than 10% of normal cells). PKR K296R expression reduced apoptosis of FA-C cells treated with IFN-␥ and TNF-␣ by more than 70% (Fig. 5B).
We evaluated further the effect of FA mutation on dsRNAinduced PKR activation by expressing the FA-C patient-derived mutant FANCC L554P cDNA in WT and PKR-null MEFs. WT cells were highly responsive to dsRNA treatments, as indicated by the dense phosphorylated (PϪ) PKR band (Fig.  5C, lane 4). Remarkably, forced expression of FANCC L554P led to detectable PKR phosphorylation in untreated WT cells (Fig. 5C, lane 5) and a significant increase of phosphorylated PKR in dsRNA-treated WT cells (lane 6). In contrast, PKR o/o cells transduced with vector alone or FANCC L554P contained neither phosphorylated (P-PKR) nor total PKR (Fig. 5C, lanes  3, 4, 7, and 8). This is expected because these cells are devoid of PKR (25). To substantiate further the idea that overexpression of FA mutant augments PKR activation, we examined the survival of cells expressing the FANCC L554P protein.
Although there was adequate expression of the FANCC L554P protein in both WT and PKR-null MEFs (Fig. 5C), the effect of the mutant FA protein on cell death was observed only in WT MEFs (Fig. 5D), suggesting a PKR-dependent effect of the mutant FA protein. Significantly, expression of the FANCC L554P protein resulted in substantial cell death in untreated WT cells (Fig. 5D). Collectively, these results support our hypothesis that FA mutations could cause abnormal PKR activation, leading to cellular apoptosis.
Functional Interaction between PKR and the FA Proteins Leads to Control of PKR Activity-To gain insight into the mechanism of molecular interaction between PKR and the FA proteins, we wished to determine whether functionally significant amounts of other components of the complex had coprecipitated with the FANCC immunocomplex. We performed PKR kinase assays using both the anti-FANCC immunopre- cipitate obtained from a normal extract and purified proteins. Fig. 6A shows the FANCA, FANCC, or FANCG proteins purified by individual FA antibody conjugated with paramagnet beads. The PKR kinase activity of the immunoprecipitate isolated by anti-FANCC beads from a normal extract was then tested in the presence of the purified FANCA, FANCC, FANCG, Hsp70, and eIF2␣ proteins. As shown in Fig. 4B, a significant amount of eIF2␣ in the reaction mixture containing only the anti-FANCC beads was phosphorylated as compared with that with both the beads and all purified components, in which a low level of phosphorylated eIF2␣ was detected (Fig.  6B, left, two upper panels, compare lane 1 with lane 4). Reactions containing eIF2␣ alone or only the purified components had no phosphorylated eIF2␣ (Fig. 6B, lanes 2 and 3), indicating that phosphorylation of eIF2␣ was dependent on PKR activation. This was confirmed by phosphorylation of the kinase (Fig. 6B, left, two lower panels). To determine the contribution of individual components to the control of PKR kinase activity, each purified component was individually omitted from the reaction mixture containing the BM immunoprecipitate (Fig.  6B, right panel). We observed significant eIF2␣ phosphorylation, about 30 -50% higher than that of the complete reaction, in the absence of either FANCA or FANCG (Fig. 6B, left,  compare lanes 5 and 7 with lane 4). On the other hand, only a marginal increase (ϳ10% higher) in phosphorylation of eIF2␣ was found when FANCC or Hsp70 was omitted (Fig. 6B, left,  compare lanes 6 and 8 with lane 4).
To verify independently the functional interaction between these components, we performed a similar assay using the anti-FANCC immunoprecipitate isolated from the extract of FANCAϪ/Ϫ lymphoblasts exposed to IFN-␥ and TNF-␣, as described in Fig. 2. The immunoprecipitate from the IFN-␥and TNF-␣-stimulated FANCAϪ/Ϫ extract had an activity to phosphorylate eIF2␣ significantly higher than that of the reconstituted reaction mixture supplemented with the complete purified components (Fig. 6C, left, compare lane 2 with lane 5). Omission of either the purified FANCA or FANCG resulted in an increased eIF2␣ phosphorylation to the level comparable with that in the reaction mixture containing the immunoprecipitate only (Fig. 6C, left, compare lanes 6 and 8 with lane 2), indicating that these two FA proteins are important for the control of PKR kinase activity. As in the previous assay with normal immunoprecipitate (Fig. 6B), removal of FANCC or Hsp70 from the complete reaction mixture had no further increase in eIF2␣ phosphorylation compared with the full reaction (Fig. 6C, left, compare lanes 7 and 9 with lane 5). Taken together, these results reveal different strength of interactions between the components of the FA-PKR signaling complex and suggest special functional significance of these interactions. DISCUSSION FA constitutes the most frequent genetic cause of BM failure (reviewed in Refs. 11 and 12). Mutations in three of the 11 FA genes, FANCA, FANCC, and FANCG, account for nearly 90% of FA patients diagnosed worldwide (13). In this report, we have used a BM extract to show that these three FA proteins functionally interact with the pro-apoptotic kinase PKR. We have also provided evidence that this interaction plays a role in the regulation of the stress kinase PKR and the downstream translational control. In primary human BM cells, disruption of the interaction between PKR and the FA proteins correlates with elevated PKR activation and hypersensitivity of BM progenitor cells to growth repression mediated by the inhibitory cytokines IFN-␥ and TNF-␣. Specific inhibition of PKR by 2-AP in these FA BM cells attenuates PKR activation and apoptosis induction by IFN-␥ and TNF-␣. In lymphoblasts derived from an FA-C patient, overexpression of a dominant negative mutant PKR (PKR K296R ) suppressed PKR activation and apoptosis in-duced by IFN-␥ and TNF-␣. Furthermore, by using genetically matched WT and PKR-null cells, we demonstrated that forced expression of a patient-derived FA-C mutant (FANCC L554P ) augmented dsRNA-induced PKR activation and cell death. We also used reconstituted kinase assay to demonstrate functional interactions between the FA proteins and PKR. Collectively, these results support our hypothesis that inappropriate activation of PKR as a consequence of certain FA mutations might play a role in bone marrow failure as frequently occurred in FA.
Our present results suggest that FANCC plays an essential role in PKR signaling in hematopoietic cells. The accumulation of PKR on the FANCC protein in FA lymphoblast cell lines (Figs. 1 and 2) and primary BM cells from FA patients (Fig. 3) suggests that FANCC binds to PKR and then recruits other components to the complex, resulting in the displacement of FANCC and control of PKR activity. The release of FANCC from the complex may depend upon a functional interaction between FANCC and the other two FA proteins (12,18) or the association of FANCC with Hsp70 (23,24). Disruption of these interactions would be expected to lead to the kind of abnormal PKR activation in FA cells (Fig. 2B and Fig. 3, C and D). In addition, overexpression of the patientderived FA-C mutant (FANCC L554P ) enhanced dsRNA-induced PKR activation and cell death (Fig. 5). A special or perhaps the central role of FANCC in PKR signaling is anticipated from the work of many laboratories on the function of FANCC that has established a consistent signaling abnormality in hematopoietic cells bearing FANCC mutations. It has been shown that suppression of FANCC expression represses clonal growth of normal erythroid and granulocytemacrophage progenitor cells, and disruption of the FANCC gene, in mice, renders hematopoietic progenitor cells hypersensitive to the pro-apoptotic effect of IFN-␥ and TNF-␣ (11, 29 -31). The multifunctional property of the FANCC protein appears to explain the clinical observations that patients in complementation group C had a significantly more severe phenotype than other complementation groups (11).
In addition to FANCC, the FANCA and FANCG proteins are necessary for the control of the PKR kinase activity. This strengthens the idea of a primary function of the two FA proteins in the FA-PKR interaction. Such a function would also explain why FANCA Ϫ/Ϫ and FANCG Ϫ/Ϫ cells accumulate abnormal FANCC⅐PKR⅐Hsp70 ternary complexes (Figs. 1-3). What might be the function of FANCA and FANCG in the PKR signaling pathway is unclear at this stage. We only recovered relatively small amounts of FANCA and FANCG proteins in human BM cells even after washing the anti-FANCC immunocomplex with a mild ionic strength buffer (Fig. 3), suggesting a weak or transient interaction of the two FA proteins with the rest of the associated proteins. However, the importance of this weak or transient interaction was demonstrated by a reconstituted kinase assay, in which omission of either FA protein resulted in a marked increase in PKR activation (Fig. 6). The existence of interactions of various strengths between the components of the FA-PKR complex is not surprising because there is evidence that tight association has been observed between FANCA and FANCG or FANCC and Hsp70 by yeast twohybrid analysis or immunoprecipitation (24,34,35), whereas weaker interactions were detected between FANCA or FANCG and FANCC by antibody methods (36,37). Thus, there are specific contacts between components of the complex, some of which may reflect contacts made between components during stress response and different steps of the complex assembly.
One objective of our study was to look for evidence that FA interacts with PKR and controls the kinase activity in hematopoietic cells. To this end, we used a combined immunoprecipitation and reconstituted kinase assay in which an active PKR kinase complex was captured from a normal extract by FANCC antibody-coated paramagnetic beads. It is noteworthy that the anti-FANCC immunocomplex can only partially suppress PKR activation, even after washing the beads with a buffer of low ionic strength (50 mM KCl). However, we were able to reconstitute the complex to nearly full strength in controlling PKR activity by adding the purified components of the complex (Fig. 6B). It appears that reconstitution of the anti-FANCC immunoprecipitate of normal cells with purified FANCA and FANCG is sufficient for full control of the PKR kinase, as addition of FANCC or Hsp70 to the reaction mixtures did not further reduce PKR activity.
The physiological relevance of the FA-PKR interaction was demonstrated by the results obtained from experiments with BM cells of the FA patients. The anti-FANCC immunoprecipitates from the FA patients are similar to those obtained with the FA mutant cell lines (Figs. 1-3). Specifically, these immunocomplexes contain significantly more FANCC-bound PKR and exhibit higher levels of the kinase activity of PKR as compared with those in the immunoprecipitate of the normal control under the same conditions. Furthermore, these biochemical alterations correlate with IFN-␥and TNF-␣-mediated growth inhibition of the BM progenitor cells of the FA-A patient (Fig. 4). We observed a 3-4-fold increase of FANCC-bound PKR in BM cell extracts of the FA-A and FA-G patients (Fig. 3) and a lymphoblast cell line that expresses a missense mutated FANCC protein (Figs. 1 and  2). This finding suggests two possible scenarios for the function of the FA proteins in PKR signaling. First, the three FA proteins bind to PKR as a pre-assembled complex in response to stresses and then recruit the PKR inhibitor to the complex. In this way, the FA proteins may function as a scaffold network. So far, no biochemical activities have been found for any of the FA proteins. However, FA proteins have been detected in at least three biologically functional complexes (18,23,38). The multifunctionality of the FA proteins may well play a putative scaffold role in multiple biological processes. In the PKR signaling pathway, for example, the FA proteins can assemble a functional signaling module and facilitate the control of the pro-apoptotic kinase mediated by a cellular PKR inhibitor. The second scenario concerns the accumulation of PKR on FANCC in extracts of BM cells from FA patients and of FA lymphoblasts. This suggests that the kinase, and perhaps together with other associated factors, remains trapped after its assembly on the FANCC protein, most likely due to a block in subsequent steps of maturation and dissociation of the complex. This strongly argues for a role of the FANCA and FANCG proteins in either the displacement of the FANCC or the transfer of PKR to a cellular PKR inhibitor. In summary, our results suggest that through interaction between FANCC and PKR, the three FA proteins can form an initial complex with the kinase, followed by the recruitment of the molecular chaperone Hsp70 to the complex. Hsp70 may exert its chaperone function by displacing the FA proteins from the initial complex and exposing the kinase to a cellular PKR inhibitor, which prevents phosphorylation of PKR and eIF-2␣ and allows protein synthesis to continue and cells to survive. In FA cells, mutations in the three FA genes cause the abnormal accumulation of PKR on the FANCC protein. As a result, when hematopoietic cells from these FA patients are exposed to environmental cues such as viral infection, growth-inhibitory cytokines (IFN-␥ and TNF-␣), or oxidative stress (e.g. reactive oxygen species), the pro-apoptotic kinase PKR becomes activated by autophosphorylation, leading to the phosphorylation of eIF-2␣ and cell death. Uncontrolled activation of PKR may contribute in part to excessive cell death, leading to the universal hallmark of FA, bone marrow failure.