ALADIN Causes Selective Failure of Nuclear Protein Import And

Home , Aladin (protein), Aprataxin, Nuclear protein

ALADINI482S causes selective failure of nuclear protein import and hypersensitivity to oxidative stress in triple A syndrome

Makito Hirano*, Yoshiko Furiya*, Hirohide Asai*, Akira Yasui†, and Satoshi Ueno*‡

*Department of Neurology, Nara Medical University, 840 Shijo-cho, Kashihara 634-8522, Japan; and †Department of Molecular Genetics, Institute of Development, Aging, and Cancer, Tohoku University, Sendai 980-8575, Japan

Edited by William S. Sly, Saint Louis University School of Medicine, St. Louis, MO, and approved December 21, 2005 (received for review July 4, 2005) Triple A syndrome is an autosomal recessive neuroendocrinological the receptor (3). Although these findings have improved our disease caused by mutations in a gene that encodes 546 amino acid understanding of the biology of nucleocytoplasmic transport, it residues. The encoded protein is the nucleoporin ALADIN, a com- remains unclear which level of disruption results in cell dysfunction ponent of nuclear pore complex (NPC). We identified a mutant leading to human disease. Cronshaw et al. (4) have recently shown ALADINI482S that fails to target NPC and investigated the conse- that most disease-associated mutant ALADINs are predominantly quences of mistargeting using cultured fibroblasts (I482Sf) from a localized in the cytoplasm and not correctly targeted to NPCs. patient with triple A syndrome. ALADINI482S affected a karyo- However, further investigations are required to clarify how mistar- pherin-␣͞␤-mediated import pathway and decreased nuclear ac- geting participates in the pathogenesis of disease. This point re- cumulations of aprataxin (APTX), a repair protein for DNA single- mains unclear primarily because ALADINQ15K causes triple A strand breaks (SSBs), and of DNA ligase I in I482Sf. This decrease syndrome in homozygotes, despite the fact that it precisely localizes was restored by wild-type ALADIN. ALADINI482S had no effect on to NPC (4). imports of M9͞kap-␤2, BIB͞kap-␤3, histone H1͞importin 7, the Recent experimental and clinical studies have shed light on the ubiquitin conjugating enzyme UbcM2͞importin 11, or the spliceo- causative relationship between neurodegeneration and defects in some protein U1A, indicating that ALADINI482S selectively impaired repair of DNA damage. Exposure of adult motor neurons in transport of discrete import complexes through NPC. Cell survival suspension to oxidizing agents can induce both DNA single-strand assay showed hypersensitivity of I482Sf to L-buthionine-(S,R)- breaks (SSBs) and neurodegeneration (5). Reactive oxygen species sulfoximine (BSO), a glutathione-depleting agent. BSO decreased (ROS) induce nuclear loss of DNA repair proteins (Ku 70 and Ku nuclear APTX and ligase I levels in I482Sf and normal control 80), leading to apoptosis in mammalian cells (6). Aprataxin fibroblasts, but increased SSBs only in I482Sf. These observations (APTX), an SSB repair protein, deserves particular attention for several reasons. APTX-defective cells are hypersensitive to geno- implied that I482Sf are hypersensitive to BSO and no longer toxic chemicals that produce SSBs (7). Mutated APTX leads to the sufficiently repair SSBs. Consistent with this notion, I482Sf trans- development of autosomal recessive cerebellar ataxia (7–9). X-ray fected with both APTX and ligase I had increased resistance to BSO, repair cross-complementing group 1 (XRCC1), a molecular scaf- whereas I482Sf transfected with LacZ vector remained hypersen- fold of the DNA repair complex for SSBs, interacts with APTX (9). sitive to BSO. We propose that oxidative stress aggravates nuclear We have focused on DNA ligase I and chromatin assembly factor import failure, which is already compromised in patient cells. 1 (CAF1) p150 protein, two factors that participate in repairing Consequent DNA damage, beyond the limited capacity of DNA ROS-induced SSBs (10). Androgen receptor (AR) regulates gene repair proteins, i.e., APTX and ligase I, may participate in triggering transcription, and its mutant causes a human neurodegenerative cell death. disease, spinal, and bulbar muscular atrophy (11). In biological and medical contexts, we decided to analyze nuclear accumulations of aprataxin ͉ DNA ligase I ͉ nuclear pore complex the above proteins (APTX, XRCC1, ligase I, CAF1, and AR) to test the ability of cells to respond to oxidative stress and to repair riple A syndrome is a neuroendocrinological disease with an DNA damage. Tautosomal recessive mode of inheritance (1). In addition to the We report a mutant ALADINI482S that failed to target NPC in clinical triad of adrenocorticotropic hormone-resistant adrenal a pair of Japanese siblings with triple A syndrome. Here we provide failure that occasionally leads to life-threatening hypoglycemia, information on the consequences of ALADIN mistargeting in achalasia, and alacrima, triple A syndrome is commonly associated patient fibroblasts, termed I482Sf. We show that selective nuclear with progressive neurological impairments involving the cranial import failure was the primary defect caused by disease-associated nerves, autonomic nervous system, and motor neurons. Causative mutant ALADIN, which increased the susceptibility of cells to mutations were recently identified in a gene that encodes 546 amino oxidative stress and led to the accumulation of damaged DNA, acid residues, possessing a molecular mass of 60 kDa (2). The leading to cell death. encoded protein, termed ALADIN, is a component of nuclear pore Results complex (NPC). ALADIN was the first nucleoporin linked to I482S human inheritable disease (2). Vertebrate NPC comprises large ALADIN in Patients. Sequence analysis of the ALADIN gene multiprotein complexes (125 MDa) consisting of 80–100 different revealed that the two siblings with triple A syndrome were ho- nucleoporins. The NPC embedded in the nuclear envelope contains an aqueous channel and acts as the sole gatekeeper controlling the Conflict of interest statement: No conflicts declared. exchange of macromolecules Ͼ40–60 kDa between the cytoplasm and nucleus (3). To overcome this barrier, proteins destined for This paper was submitted directly (Track II) to the PNAS office. nuclear import have specific signals that are recognized by a Abbreviations: NPC, nuclear pore complex; NLS, nuclear localization signal; SSB, single- strand break; XRCC1, X-ray repair cross-complementing group 1; CAF1, chromatin assembly structurally related family of soluble transport receptor proteins, factor 1; APTX, aprataxin; AR, androgen receptor; GFPx2, GFP–GFP; BSO, L-buthionine- collectively termed karyopherin (kap). Typically, nuclear localiza- (S,R)-sulfoximine; GSH, glutathione; DEC, decylubiquinone; VA, all-transretinol; ROS, re- tion signal (NLS) containing cargo is bound by transport receptors. active oxygen species; 8-oxoG, 7,8-dihydro-8-oxoguanine. After the cargo-transport receptor complex is translocated into the ‡To whom correspondence should be addressed. E-mail: [email protected]. nucleus through NPC, RanGTP induces release of the cargo from © 2006 by The National Academy of Sciences of the USA

2298–2303 ͉ PNAS ͉ February 14, 2006 ͉ vol. 103 ͉ no. 7 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505598103 Downloaded by guest on September 27, 2021 Fig. 1. ALADINI482S mislocalizes from nuclear pore complex (NPC) to cytoplasm in transfected HeLa cells. The localization of GFP–ALADIN is shown in the left column (green), and NPCs immunostained with monoclonal antibody mAb414 are shown in the middle column (red). The distribution of wild-type (wt) ALADIN and ALADINQ15K overlapped with that of NPC, whereas ALADINI482S and ALADIN312del localized to the cytoplasm, rather than to NPC in the right column (green͞red).

mozygous and their mother was heterozygous for a T-to-G trans- MEDICAL SCIENCES version in exon 16, resulting in a substitution from phylogenetically conserved isoleucine to serine at codon 482, ALADINI482S. This mutation was identified in blood DNA samples from the three family members, as well as in skin fibroblast DNA from the female Fig. 2. ALADINI482S affects nuclear import pathways in I482S fibroblasts patient. I482Sf and age- and sex-matched normal control fibroblasts (I482Sf). (A) The kap-␣͞␤-mediated nuclear imports of GFPx2–NLSAPTX and were primarily cultured for this study. GFPx2–NLSSV40 were decreased in I482Sf. GFPx2–NLSXRCC1(239–266) showed identical distributions in both I482Sf and normal control fibroblasts. stNLSXRCC1(239– ALADINI482S Mislocalized from NPC to Cytoplasm. HeLa cells were 403) was highly efficient for nuclear import in both I482Sf and normal control ͞ ␤ ͞ ␤ transfected with GFP–wild-type (wt) and GFP–mutant ALADINs fibroblasts. I482Sf showed normal nuclear imports of M9 kap- 2, BIB kap- 3, histone H1͞importin 7, and UbcM2͞importin 11 complexes and the spliceo- and then fixed and stained with monoclonal antibody mAb414 as an some protein U1A. (B) The above data expressed in a bar graph format NPC marker. mAb414 recognizes nucleoporins (Nup 358, Nup214, (mean Ϯ SD). *, P Ͻ 0.05, unpaired Student’s t test as compared with control Nup153, and Nup62) located at the cytoplasmic or nuclear faces of cells. NPC. GFP–wtALADIN and GFP–ALADINQ15K localized to NPC, whereas GFP–ALADINI482S and GFP–ALADIN312del mislocalized to the cytoplasm (Fig. 1). Although the findings for The nuclear import of GFPx2-NLSXRCC1(239–266) (mediated by ALADINI482S were consistent with those for most other ALADIN kap-␣͞␤) was identical in I482Sf and control cells. An additional mutants as reported by Cronshaw et al. (4), ALADINI482S was truncation study revealed that a stretched sequence termed st- identified between 478 and 499 aa, supporting the notion that these NLSXRCC1(239–403) very efficiently allows GFPx2-cargo to enter 22 aa are essential for ALADIN targeting to NPC. ALADIN312del, nuclei in both types of cells. These observations showed that the a misspliced fragment (lacking exons 10 and 11) encoding the nuclear import pathway for XRCC1 was not affected by N-terminal 312-aa region, was deduced from an RT-PCR product ALADINI482S. Nuclear import of M9͞kap-␤2, BIB͞kap-␤3, histone derived from normal human lymphocyte mRNA. This truncated H1͞importin 7, and the ubiquitin conjugating enzyme UbcM2͞ ALADIN312del localized predominantly to cytoplasm, but also importin 11 did not discernibly differ between I482Sf and control clearly entered nuclei, suggesting that the N-terminal 312-aa region cells. Nuclear import of the spliceosome protein U1A, which is alone was not sufficient and that an additional C-terminal region independent of cytosolic-transport factors, was also similar in both was required for precise targeting and retention of ALADIN at types of cells (Fig. 2). NPC. The immunostaining patterns of mAb414-positive nucleoporins, lamin B, and kap-␤ were identical to those of I482Sf and ALADINI482S Decreases Nuclear Accumulation of Specific Proteins. normal control fibroblasts, indicating that ALADIN mistargeting Fluorescent microscopic measurement revealed that nuclear GFP– did not disrupt NPC structure (data not shown). APTX and GFP–ligase I levels were reduced in I482Sf, as compared with those in normal control fibroblasts (Fig. 3). Dose- ALADINI482S Affects a kap-␣͞␤-Mediated Import Pathway. We mon- dependent transfection confirmed that I482Sf had impaired nuclear itored the fluorescence intensity of GFP–GFP (GFPx2)–NLS con- import of APTX (Fig. 3A Right). In contrast, GFP–XRCC1, structs representing known cargoes from defined import pathways GFP–CAF1, and GFP–AR showed similar distribution patterns for in I482Sf and normal control fibroblasts. The nuclear accumulation I482Sf and control cells (Fig. 3B). These findings supported the idea of cargo was estimated as the fluorescence intensity in the nucleus that ALADINI482S selectively affects transport of discrete import to that in the cell and was expressed as a percentage. The nuclear complexes through NPC. The possibility that test proteins simply levels of GFPx2-NLSAPTX and GFPx2-NLSSV40 (both are mediated diffused through NPC was excluded for several reasons. Estimated by kap-␣͞␤) were significantly lower in I482Sf than in control cells. molecule sizes of GFPs tested here exceed the NPC passive

Hirano et al. PNAS ͉ February 14, 2006 ͉ vol. 103 ͉ no. 7 ͉ 2299 Downloaded by guest on September 27, 2021 Fig. 3. Nuclear import of GFP–classic NLS bearing nucleoprotein in I482S fibroblasts (I482Sf) and normal control fibroblasts. (A) GFP–APTX was localized to the Hoechst dye 33258-stained nuclei in control cells, but was localized to the cytosol and nuclei in I482Sf. The first column is shown at low magnifi- cation (open bar, 10 ␮m). Reduced nuclear APTX levels in I482Sf as compared with control cells was confirmed after transfection with 1–4 ␮g of plasmid DNA containing APTX cDNA (mean Ϯ SD; *, P Ͻ 0.05, unpaired Student’s t test). (B) GFP–ligase I localized to both cytoplasm and nuclei in I482Sf, but only to nuclei in control cells. GFP–XRCC1, GFP–CAF1, and GFP–AR showed similar distribution patterns in both types of cells. The distributions are expressed in a bar graph format (mean Ϯ SD). *, P Ͻ 0.05, unpaired Student’s t test as compared with control cells.

Fig. 4. Complementation of nuclear import failure in I482S ﬁbroblasts diffusion limit of 40–60 kDa (3). Monitoring of nuclear protein (I482Sf). (A) The decreased nuclear accumulations in I482Sf were normalized accumulations in unfixed cells showed that nuclear import occurred by coexpression of GFP–wtALADIN, partly restored by GFP–ALADINQ15K, and against a concentration gradient. We confirmed the reduction in not changed by GFP–ALADINI482S or by GFP–ALADIN312del.(B) The above data Ϯ Ͻ nuclear import of APTX and ligase I by experiments using myc- expressed in a bar graph format (mean SD). *, P 0.05, unpaired Student’s t test as compared with control cells; #, P Ͻ 0.05, as compared with I482Sf proteins (myc-APTX and myc-ligase I) (Fig. 4). The reduction in transfected with myc-APTX or myc-ligase I alone [GFP–ALADIN(Ϫ)]. nuclear import of APTX and ligase I was attributable to ALADINI482S, because it was restored by coexpression of GFP– wtALADIN, but not by that of GFP–ALADINI482S or GFP– control-cell survival (Fig. 5A). The antioxidants decylubiquinone ALADIN312del. The reduction in nuclear import of APTX and (DEC) and all-transretinol (VA) inhibited BSO-induced cell death ligase I was not augmented by coexpression of ALADINI482S, (Fig. 5B). The cell death at 0.1 mM BSO treatment for 48 h was also excluding the possibility that ALADINI482S, lacking NPC targeting significantly reduced by transfection of GFP–wtALADIN (81.3 Ϯ ability, bound to the kap-cargo and that resulting trimetric com- 5.0%, mean Ϯ SD), but not by that of GFP–ALADINI482S (51.4 Ϯ plexes (ALADINI482S-kap-APTX or ligase I) were retained within 1.5%) or LacZ gene (50.0 Ϯ 3.2%). I482Sf and normal control the cytoplasm, thereby decreasing nuclear protein import in I482Sf. fibroblasts, after 14-hr incubation with 0.1 and 1 mM BSO, were Of interest is the point mutation Q15K, which precisely localizes to assayed for GSH and highly reactive oxygen species (hROS). These NPC, but only partially restores the decreased nuclear accumula- conditions did not cause cell death in either type of cells. Total tions of APTX and ligase I in I482Sf, indicating that ALADINЈs cellular GSH levels estimated from ratios of GSH in BSO-treated targeting to NPC is essential but not sufficient for normal NPC cells͞GSH in nontreated cells were similar in I482Sf [3.9 Ϯ 0.2% functions (Fig. 4). (mean Ϯ SD) of base line] and control cells (4.2 Ϯ 2.5%). Cellular hROS increased to very similar levels in both types of cells. DEC I482S Fibroblasts Are Hypersensitive to Oxidative Stress. We treated inhibited this BSO-induced elevation of cellular hROS levels (Fig. I482Sf with L-buthionine-(S,R)-sulfoximine (BSO) to analyze the 5C). DNA fragmentation, characteristic of apoptotic cell death, in consequences of the decreased nuclear levels of APTX and ligase I482Sf was not apparent until 24 h after 1 mM BSO treatment, but I. BSO is a highly selective inhibitor of ␥-glutamylcysteine syn- was obvious at 48 h (Fig. 5D). Serial nuclear staining with Hoechst thetase, and its addition to culture medium effectively arrests dye 33258 showed increasing populations of cells with apoptotic glutathione (GSH) synthesis. In this model of oxidative stress, the chromatin condensation, accounting for 0%, Ϸ20%, and Ͼ80% of endogenous formation of ROS is largely unopposed, consequently I482Sf after 14, 24, and 48 h of BSO treatment, respectively (Fig. resulting in oxidative cell damage. Treatment of cells for 48 h with 5E). DEC and VA efficiently protected BSO-treated cells from 1 mM BSO decreased the viability of I482Sf to Ϸ20% of the DNA fragmentation and chromatin condensation (Fig. 5 D and E, base-line level in a dose-dependent manner, but did not affect respectively). These findings indicated that the effects observed

2300 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505598103 Hirano et al. Downloaded by guest on September 27, 2021 MEDICAL SCIENCES

Fig. 5. Increased sensitivity of I482S fibroblasts (I482Sf) to oxidative stress. (A) Survival of control and I482Sf after treatment with BSO in a dose- dependent manner. The bar represents the SD from five independent experiments. *, P Ͻ 0.01, Mann–Whitney U test as compared with control cells. (B) Antioxidant treatment with decylubiquinone (DEC) or VA inhibited BSO- induced cell death. The bar represents the mean Ϯ SD from five independent experiments. *, P Ͻ 0.05, Mann–Whitney U test as compared with control cells. (C) BSO treatment for 14 h increased highly reactive oxygen species (hROS) to Fig. 6. Targeting APTX and ligase I into nucleus increased resistance of I482S a similar level in control cells and I482Sf, without cell death. Treatment with fibroblasts (I482Sf) to oxidative stress. (A) Reduced nuclear accumulation of 1 ␮M DEC reduced the fluorescence intensity in both types of cells. #, P Ͻ 0.05, GFP–APTX by L-buthionine-(S,R)-sulfoximine (BSO) treatment for 14 h in both unpaired Student’s t test as compared with nontreated cells; n.s., nonsignif- I482Sf and normal control fibroblasts, without cell death. The reduction was icant as compared with control cells. (D) Time course analysis of DNA frag- restored by 1 ␮M decylubiquinone (DEC) treatment. The bar represents the SD. mentation. Treatment of I482S fibroblasts with 1 mM BSO for 48 h induced , P Ͻ 0.05, unpaired Student’s t test as compared with control cells; #, P Ͻ 0.05, DNA fragmentation, which was inhibited by 1 ␮M DEC or 5 ␮M VA treatment. * as compared with nontreated cells. (B) Comet assay for I482Sf treated with DNA from pheochromocytoma PC12 cells that underwent serum deprivation- BSO for 14 h, without cell death or apoptotic DNA fragmentation. Higher induced apoptosis was used for positive control (PC). M, 123-bp ladder marker. levels of DNA strand breaks (tail moment), mostly reflecting single strand (E) Fluorescent images of fibroblast nuclei stained with Hoechst dye 33258. breaks, were detected in I482Sf, as compared with control cells. The bar Treatment of I482S fibroblasts with 1 mM BSO for 48 h induced chromatin represents the mean Ϯ SD from five independent experiments. , P Ͻ 0.05, condensation, which was reduced by 1 ␮M DEC or 5 ␮M VA treatment. * Mann–Whitney U test as compared with control cells. (C) Immunostaining of 7,8-dihydro-8-oxoguanine in cells treated with BSO for 14 h, without cell death. Stronger nuclear immunoreactivity was observed in I482Sf than in after treatment with BSO, which may be ascribed to GSH defi- control cells. (D) Restoration of reduced nuclear accumulations of GFP–APTX ciency, were caused by ROS. Taken together, the above studies and GFP–ligase I in I482Sf by attaching stNLSXRCC1(239–403) (stNLS). GFP–stNLS- I482S showed that ALADIN increased the susceptibility of I482Sf to APTX and GFP–stNLS-ligase I were localized to the nucleus. (E) 14-h BSO oxidative stress. treatment did not impair nuclear import of GFP–stNLS-APTX or GFP–stNLS- ligase I in I482Sf or control cells. The bar represents the mean Ϯ SD. (F) DNA Repair Proteins, DNA Damage, and Cell Death of I482S Fibro- Complementation of BSO sensitivity. Cell death was decreased by the expres- blasts. After incubation with 0–1 mM BSO for 14 h, the nuclear sion of GFP–wtALADIN and by the coexpression of GFP–stNLS–APTX and level of APTX decreased in a dose-dependent manner to 20% in GFP–stNLS–ligase I, but not by the expression of APTX or ligase I alone. The bar represents the SD from five independent experiments. *, P Ͻ 0.05, Mann– I482Sf and 40% in normal control fibroblasts. These reductions Whitney U test as compared with cells transfected with LacZ vector. were restored by DEC (Fig. 6A). Comet assay conditions included neutralized electrophoresis for double-strand breaks (DSBs) and alkaline rewinding coupled with neutralized electrophoresis for both SSBs and DSBs. I482Sf accumulated higher levels of SSBs 8-oxoG in cells treated with BSO revealed stronger nuclear immu- than control. DSBs did not exceed the reliable minimum detection noreactivity in I482Sf, indicating defects in the repair pathway for level of our system (tail moment ϭ 1) in I482Sf or in control cells 8-oxoG, related to base excision repair together with SSB repair (Fig. 6B). Increased ROS produce SSBs, and also generate oxida- (Fig. 6C). On the basis of these observations, we proposed that tively modified DNA base products, including 7,8-dihydro-8- decreased levels of intranuclear APTX and ligase I in I482Sf no oxoguanine (8-oxoG) as a major modified base. Immunostaining of longer sufficiently repaired damaged DNA.

Hirano et al. PNAS ͉ February 14, 2006 ͉ vol. 103 ͉ no. 7 ͉ 2301 Downloaded by guest on September 27, 2021 Complementation of I482S Fibroblast Hypersensitivity to Oxidative facilitate the import of nuclear proteins, including APTX and ligase Stress. Our results suggested a plausible relation between the I. Notwithstanding the complexity involved, inherent to nucleocy- decreased levels of nuclear SSB repair proteins (APTX and ligase toplasmic transport machinery, we provided in vitro evidence I) and the hypersensitivity of I482Sf to oxidative stress, leading to suggesting that ALADINI482S affects the ability of cells to respond cell death. To confirm this linkage, we constructed a vector that can to oxidative stress and repair damaged DNA. An important ques- enter nuclei via an import pathway not affected by ALADINI482S. tion to be addressed is why cotransfection with APTX and ligase I The vector contains cDNA coding for GFP–stNLSXRCC1(269–403)– is required for I482Sf to increase resistance against oxidative stress. APTX or GFP–stNLSXRCC1(269–403)–ligase I. stNLSXRCC1(269–403) APTX was reported to interact with DNA repair proteins such as proved to be highly effective for the nuclear import of APTX and XRCC1 (9). We previously showed that XRCC1 plays an important ligase I in I482Sf and normal control fibroblasts cultured in media ‘‘scaffolding’’ role in organizing SSB repair and participates in containing 0–1.0 mM BSO (Fig. 6 D and E). Coexpression of ‘‘long-patch repair’’ pathway in which ligase I functions (10). GFP–stNLSXRCC1(269–403)–APTX and GFP–stNLSXRCC1(269–403)– Consistent with these notions, cooperative DNA repair by APTX ligase I increased the viability of I482Sf treated with 0.1–1.0 mM and ligase I increased the resistance of I482Sf to oxidative stress. BSO for 48 h. Notably, however, neither APTX nor ligase I alone Collectively, we propose the following hypothetical mechanism for could rescue BSO-treated I482Sf from cell death. I482Sf trans- triple A syndrome. The major affected organs in this disease, the fected with LacZ vector remained hypersensitive to BSO (Fig. 6F). adrenal and nervous systems, have high requirements for ATP and These results support the correlation of the ALADINI482S- are therefore exposed to high amounts of ROS generated during dependent failure of nuclear import for APTX and ligase I with the mitochondrial oxidative phosphorylation. Oxidative cell stress aug- hypersensitivity of I482Sf to ROS. ments nuclear import failure, which is already compromised. The resulting DNA damage, which is beyond the limited capacity of Discussion DNA repair proteins, i.e., APTX and ligase I, may participate in We showed that I482Sf obtained from a patient with triple A triggering cell death. syndrome offered an experimental system for the study of nuclear In conclusion, our results provide important information on the protein import under oxidative stress. Here, we provide in vitro consequences of ALADIN mistargeting in triple A syndrome. We evidence supporting the notion that I482Sf are defective in the showed that selective nuclear import failure was the primary defect nuclear import of APTX and ligase I, and that ALADINI482S affects caused by disease-associated mutant ALADIN, which increased the the ability of cells to respond to oxidative stress and to repair susceptibility of cells to oxidative stress and accumulated damaged damaged DNA, leading to cell death. DNA, culminating in cell death. Future studies should determine which nuclear proteins are affected by different types of mutated I482S Fibroblasts Show Selective Failure of Nuclear Protein Import. ALADIN. For instance, identification of nuclear proteins affected ALADINI482S mislocalized predominantly to cytoplasm and is by NPC-targeting mutants would be of particular interest, because presumably assembled inefficiently into NPC consisting of 80–100 the Q15K mutation may be associated with milder symptoms or different nucleoporins. This ALADIN-defective NPC had a signif- later onset of disease (16). Delineation of the mutational basis for icantly impaired kap-␣͞␤-mediated import pathway and showed differences in nuclear import of proteins may help to explain the weak nuclear import of APTX and ligase I. Notably, the nuclear functions of nucleoporins at NPC, as well as the mechanisms imports of XRCC1, CAF1, and AR, each carrying kap-␣͞␤- responsible for the extreme variations in the clinical phenotypes of recognizing NLS, were not affected in I482Sf. These findings triple A syndrome. Improved understanding of the pathogenesis of suggested that I482Sf have at least two subsets of kap-␣͞␤-mediated this disease will hopefully lead to the development of new thera- pathways. Importantly ALADINQ15K only partially corrected the peutic strategies. decreases in nuclear APTX and ligase I, although it localized to NPC in I482Sf. Given that ALADINQ15K causes symptoms of triple Materials and Methods A syndrome in homozygotes, our findings suggest that nuclear Patients and Cell Culture. A 48-year-old Japanese woman and her import failure of specific proteins has a primary role in the 43-year-old brother had been surgically treated for achalasia at 12 pathogenesis of triple A syndrome. We found that the nuclear and 26 years of age, respectively. Their parents were first cousins import of cargo–receptor complexes, namely M9͞kap-␤2, BIB͞kap- and apparently had no symptoms. Clinical and laboratory tests and ␤3, histone H1͞importin 7, and UbcM2͞importin 11 complexes electrophysiological studies revealed adrenal insufficiency, reduced (12–14), and the spliceosome protein U1A showed no discernable tear production, and dysfunction of motor neurons in both patients. difference between I482Sf and normal control fibroblasts (15). On DNA was extracted from blood samples obtained from the three the basis of these results, we concluded that ALADINI482S selec- family members and subjected to DNA sequence analysis as tively impaired discrete protein-import pathways rather than pro- described (2). Fibroblasts obtained by skin biopsy of the female tein import in general. patient and normal control fibroblasts were maintained in a 5% CO2 humid atmosphere at 37°C in DMEM supplemented with 10% I482S Fibroblasts Are Hypersensitive to Oxidative Stress. Our obser- FBS. The patients and their mother gave written consent to this vations showed that I482Sf were hypersensitive to oxidative stress study. and accumulated high levels of SSBs. The link between increased SSBs and decreased APTX and ligase I levels within I482Sf nuclei Vectors, Antibodies, and Chemicals. pCMV-Myc, pEGFP–C1, and implied that cells exposed to BSO could no longer adequately repair pEGFP–N1 were purchased from Clontech; pcDNA3.1-LacZ, Li- DNA damage, leading to cell death. To confirm the importance of pofectamine 2000, and ␤-galactosidase assay kit were from Invitro- APTX and ligase I in cellular resistance to oxidative stress, we gen; and SuperFect was from Qiagen (Valencia, CA). Antibodies targeted these two DNA repair proteins into the nuclei via an to nucleoporins (mAb414), lamin B, kap-␤, and 8-oxoG were import pathway unaffected by ALADINI482S. This nuclear targeting purchased from Babco (Richmond, CA), EMD Biosciences (San was accomplished by using specifically constructed vectors carrying Diego, CA), BD Biosciences (San Jose, CA), and the Japan highly effective stNLSXRCC1(239–403). Transfection with wtALADIN Institute for the Control of Aging (Shizuoka, Japan), respectively. had a greater effect on I482Sf survival than cotransfection with both GSH quantification kit was from Dojindo Molecular Technologies APTX and ligase I did, which may be explained as follows. (Gaithersburg, MD), and aminophenyl fluorescein (APF) was from ALADINI482S may result in specific mislocalization of other nucleo- Daiichi Pure Chemicals (Tokyo). CometAssay kit was from Trevi- porins or transport factors. Transfection of I482Sf with wtALADIN gen (Gaithersburg, MD). All other chemicals and fluorescence- may correct the localization of such proteins, and as a consequence labeled secondary antibodies were obtained from Sigma.

2302 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0505598103 Hirano et al. Downloaded by guest on September 27, 2021 Vector Construction. ALADIN cDNA was PCR amplified from Analysis of I482S Fibroblast Susceptibility to BSO. I482Sf and normal reverse-transcribed RNA of control lymphocytes. ALADINI482S control fibroblasts were treated with 0–1 mM BSO for 48 h, and and ALADINQ15K cDNAs were synthesized by PCR mutagenesis. subjected to trypan blue exclusion assay to assess cell survival. The ALADIN cDNA clones were inserted into pCMV-Myc or into Intracellular levels of GSH and hROS were estimated from ratios pEGFP–C1. GFP–human AR, GFP–DNA ligase I, and GFPx2- of GSH (hROS) in BSO-treated cells (0.1 and 1 mM BSO for UbcM2 cDNAs were kindly provided by Y. Nagai, H. Leonhardt, 14 h)͞GSH (hROS) in nontreated cells. APF, a fluorescent sub- and S. Plafker (13, 17). Plasmids for GFP–APTX, myc-APTX, strate, was used to specifically quantify amounts of hROS. The myc-ligase I, GFP–CAF1, and GFP–XRCC1 were synthesized as fluorescent intensity of at least 300 cells in three independent described (8, 10, 18). To test NLS-mediated import pathway, experiments was measured with a fluorescence microscope pEGFP–N1 vectors were constructed to contain cDNA sequences (DMIRB, Leica). DNA fragmentation electrophoresis was per- encoding GFP–NLS. For instance, GFP–NLSsv40 represents a formed on 1 mM BSO-treated cells for 48 h. Cells were incubated fusion protein of NLS (PKKKRKVED) of the simian virus 40 large in the presence of 1 mM BSO for 14, 24, and 48 h and stained with T antigen tagged with GFP. NLSs were deduced from HRKRKR Hoechst dye 33258 for chromatin condensation analysis. Parallel for APTX and QESPKGKRKLDLNQEEKKTPSKPPAQLS for study conditions included 0–1 ␮M DEC, a coenzyme Q10-derivative XRCC1. A truncation study to define NLS of XRCC1 was per- antioxidant, and 0–5 ␮M VA. DNA damage was also estimated for formed as described (18). To construct GFPx2-NLS vector, each I482Sf and control cells. GFP–NLS cDNA was further inserted into modified pEGFP–N1 vector lacking the first ATG for GFP. cDNAs encoding GFP- Analysis of Transfected I482Sf Fibroblasts in Response to BSO. I482Sf nonclassical NLS, such as M9 domain of hnRNP A1 (codon and normal control fibroblasts were transfected with cDNAs 264–305), BIB domain of ribosomal protein L23a (codon 32–74), encoding GFP–APTX and incubated for 24 h. The nuclear histone H1, and splicesosome protein U1A (codon 94–204) were accumulation of GFP–APTX was monitored in both types of inserted into GFPx2 vector (12, 13, 15, 19). cells after treatment with 0–1 mM BSO for 14 h. We constructed specific vectors by attaching GFP–stNLSRCC1(239–403) (stNLS) to Vector Transfection and Monitoring of Nuclear Accumulation. The APTX and to ligase I, thereby targeting APTX and ligase I into ͞ GFP and or myc plasmid vectors were cotransfected with the nucleus. The nuclear accumulation of GFP–stNLS–APTX pcDNA3.1-LacZ vector into HeLa cells by using SuperFect or into and GFP–stNLS–ligase I in I482Sf was measured by using a I482Sf and normal control fibroblasts by using Lipofectamine 2000. fluorescence microscope. To determine the effect of nuclear- All plasmid DNA amounts for nuclear proteins (1–4 ␮g) were

imported APTX and ligase I on the sensitivity of I482Sf to BSO, MEDICAL SCIENCES ␮ adjusted by using additional LacZ vector (0.5–3 g) to maximize I482Sf were transfected with one or multiple vectors containing nuclear accumulation in control cells. Forty-eight hours after GFP–wtALADIN, GFP–stNLS–APTX, GFP–stNLS–ligase I, transfection, cells underwent fluorescent or confocal microscopic or LacZ. After treatment with 0–1 mM BSO for 48 h, the analysis. Cells expressing AR were assessed 16 h after adding 10 nM transfected I482Sf cells were subjected to trypan blue exclusion 5-dihydrotestosterone (DHT). GFPs expressed in cultured cells assay to assess cell survival. were quantified with a fluorescence microscope. Nuclear accumu- ͞ lation of GFP was defined as the ratio of nucleus cell fluorescence Assay for DNA Damage. Comet assay. DNA-strand breaks were intensity. The nucleus and the entire cell were delineated by assessed by using a CometAssay kit according to the manufac- superimposing a GFP image with a Hoechst dye 33258-stained turer’s protocol and reported methods (20). image and a phase-contrast image, respectively. myc-APTX and 8-oxoG assay. I482Sf and normal control fibroblasts (1 ϫ 104 cells per myc-ligase I detected in fixed cells by the fluorescent secondary well) were seeded on an eight-well chamber slide and cultured antibody were analyzed similarly. Data were obtained from at least overnight. After 14 h of treatment with 0–1 mM BSO, cells were 300 cells in three independent experiments. Immediately after immunostained with anti-8-oxoG monoclonal antibody according quantification, the cells were subjected to the ␤-galactosidase assay to the protocol in the literature (10) and analyzed with a fluores- to validate the equivalency level of transfection efficiency in each cence microscope. Data were obtained from at least 100 cells in experiment. three independent experiments. Immunocytochemistry. HeLa cells transfected with expression vec- GFP–human AR, GFP–DNA ligase I, and GFPx2-UbcM2 cDNAs were tors for GFP–ALADINs were immunostained with mAb414 and gifts from Drs. Y. Nagai (Osaka University, Osaka), H. Leonhardt fibroblasts were immunostained with mAb414 and antibodies (Ludwig Maximilians University, Munich), and S. Plafker (University of against nucleoporins, lamin B, or kap-␤, as described (14). Fluo- Oklahoma, Norman). This study was partly supported by Grants-in-Aids rescently labeled secondary antibodies were imaged with an MRC- for Scientific Research from the Ministry of Education, Culture, Sports, 600 laser-scanning confocal microscope (Bio-Rad). Science and Technology of Japan.

1. Allgrove, J., Clayden, G. S., Grant, D. B. & Macaulay, J. C. (1978) Lancet 1, 1284–1286. 11. Fischbeck, K. H. (1997) J. Inherited Metab. Dis. 20, 152–158. 2. Tullio-Pelet, A., Salomon, R., Hadj-Rabia, S., Mugnier, C., de Laet, M. H., 12. Jakel, S., Albig, W., Kutay, U., Bischoff, F. R., Schwamborn, K., Doenecke, D. & Chaouachi, B., Bakiri, F., Brottier, P., Cattolico, L., Penet, C., et al. (2000) Nat. Genet. Gorlich, D. (1999) EMBO J. 18, 2411–2423. 26, 332–335. 13. Plafker, S. M. & Macara, I. G. (2000) EMBO J. 19, 5502–5513. 3. Nigg, E. A. (1997) Nature 386, 779–787. 14. Wu, X., Kasper, L. H., Mantcheva, R. T., Mantchev, G. T., Springett, M. J. & van 4. Cronshaw, J. M. & Matunis, M. J. (2003) Proc. Natl. Acad. Sci. USA 100, 5823–5827. Deursen, J. M. (2001) Proc. Natl. Acad. Sci. USA 98, 3191–3196. 5. Liu, Z. & Martin, L. J. (2001) J. Comp. Neurol. 432, 35–60. 15. Hetzer, M. & Mattaj, I. W. (2000) J. Cell Biol. 148, 293–303. 278, 6. Song, J. Y., Lim, J. W., Kim, H., Morio, T. & Kim, K. H. (2003) J. Biol. Chem. 16. Sandrini, F., Farmakidis, C., Kirschner, L. S., Wu, S. M., Tullio-Pelet, A., Lyonnet, 36676–36687. S., Metzger, D. L., Bourdony, C. J., Tiosano, D., Chan, W. Y., et al. (2001) J. Clin. 7. Gueven, N., Becherel, O. J., Kijas, A. W., Chen, P., Howe, O., Rudolph, J. H., Gatti, Endocrinol. Metab. 86, 5433–5437. R., Date, H., Onodera, O., Taucher-Scholz, G., et al. (2004) Hum. Mol. Genet. 13, 1081–1093. 17. Cardoso, M. C., Joseph, C., Rahn, H. P., Reusch, R., Nadal-Ginard, B. & Leonhardt, 8. Hirano, M., Furiya, Y., Kariya, S., Nishiwaki, T. & Ueno, S. (2004) Biochem. Biophys. H. (1997) J. Cell Biol. 139, 579–587. Res. Commun. 322, 380–386. 18. Okano, S., Lan, L., Caldecott, K. W., Mori, T. & Yasui, A. (2003) Mol. Cell. Biol. 23, 9. Sano, Y., Date, H., Igarashi, S., Onodera, O., Oyake, M., Takahashi, T., Hayashi, S., 3974–3981. Morimatsu, M., Takahashi, H., Makifuchi, T., et al. (2004) Ann. Neurol. 55, 241–249. 19. Siomi, H. & Dreyfuss, G. (1995) J. Cell Biol. 129, 551–560. 10. Lan, L., Nakajima, S., Oohata, Y., Takao, M., Okano, S., Masutani, M., Wilson, S. H. 20. Singh, N. P., Tice, R. R., Stephens, R. E. & Schneider, E. L. (1991) Mutat. Res. 252, & Yasui, A. (2004) Proc. Natl. Acad. Sci. USA 101, 13738–13743. 289–296.

Hirano et al. PNAS ͉ February 14, 2006 ͉ vol. 103 ͉ no. 7 ͉ 2303 Downloaded by guest on September 27, 2021