a 127-kda protein (uv-ddb) binds to the cytoplasmic domain of the alzheimer's amyloid precursor...

A 127-kDa Protein (UV-DDB) Binds to the CytoplasmicDomain of the Alzheimer’s Amyloid Precursor Protein

Takuo Watanabe, *Jun Sukegawa, *Izumi Sukegawa, †Susumu Tomita, †Ko-ichi Iijima,†Shinobu Oguchi, †Toshiharu Suzuki, Angus C. Nairn, and Paul Greengard

Laboratory of Molecular and Cellular Neuroscience, Rockefeller University, New York, New York, U.S.A.; and*Department ofPharmacology, Tohoku University School of Medicine, Sendai; and†Laboratory of Neurobiophysics, School of Pharmaceutical

Sciences, University of Tokyo, Tokyo, Japan

Abstract: Alzheimer amyloid precursor protein (APP) isan integral membrane protein with a short cytoplasmicdomain of 47 amino acids. It is hoped that identificationof proteins that interact with the cytoplasmic domain willprovide new insights into the physiological function ofAPP and, in turn, into the pathogenesis of Alzheimer’sdisease. To identify proteins that interact with the cyto-plasmic domain of APP, we employed affinity chroma-tography using an immobilized synthetic peptide corre-sponding to residues 645–694 of APP695 and identified aprotein of ;130 kDa in rat brain cytosol. Amino acidsequencing of the protein revealed the protein to be a rathomologue of monkey UV-DDB (UV-damaged DNA-binding protein, calculated molecular mass of 127 kDa).UV-DDB/p127 co-immunoprecipitated with APP using ananti-APP antibody from PC12 cell lysates. APP also co-immunoprecipitated with UV-DDB/p127 using an anti-UV-DDB/p127 antibody. These results indicate that UV-DDB/p127, which is present in the cytosolic fraction,forms a complex with APP through its cytoplasmic do-main. In vitro binding experiments using a glutathioneS-transferase–APP cytoplasmic domain fusion proteinand several mutants indicated that the YENPTY motifwithin the APP cytoplasmic domain, which is important inthe internalization of APP and amyloid b protein secre-tion, may be involved in the interaction between UV-DDB/p127 and APP. Key Words: Amyloid precursor protein—UV-damaged DNA binding protein—DNA repair.J. Neurochem. 72, 549–556 (1999).

It is believed that mismetabolism of the Alzheimeramyloid precursor protein (APP), which leads to over-production of amyloidb protein (Ab), is central to theetiology of some forms of Alzheimer’s disease (AD)(Younkin, 1995; Selkoe, 1996). The short cytoplasmicdomain of APP (APPc) is highly conserved betweenspecies, suggesting an important function for this regionof the protein. Endocytosis of APP, which has beenshown to be involved in the formation of Ab, is mediatedby sequences in APPc (Koo and Squazzo, 1994; Lai etal., 1995). It has also been suggested that APP may

function in some aspect of signal transduction, possiblyby acting as a cell surface receptor (Neve, 1996). As thecytoplasmic domain appears to lack any enzymatic ac-tivity, it is likely that one or more proteins must interactwith the domain and that such protein–protein interac-tions will be critical to the regulation of APP function ormetabolism. In turn, a fundamental understanding ofAPP function and metabolism may help in the develop-ment of potential therapeutic agents that will be useful ininhibiting Ab formation.

Here we report identification of a protein, the 127-kDasubunit of UV-damaged DNA binding protein (UV-DDB), which has been hypothesized to be involved inDNA repair machinery, as a protein that binds to APPc.

EXPERIMENTAL PROCEDURES

Synthesis of peptidesPeptides used for preparation of the affinity column and for

antibody production were synthesized and purified by reversed-phase column HPLC (W. M. Keck Foundation BiotechnologyResource Laboratory, Yale University). The peptides had theexpected amino acid compositions and mass spectra.

Affinity chromatography of rat brain homogenatesSynthetic peptide (8 mg) encompassing APPc (residues

645–694 in APP695) was dissolved in distilled water and cou-pled to 2 ml of AminoLink Coupling Gel (Pierce) with 50 mMNaCNBH3 for 6 h atroom temperature (APPc affinity column).

Received July 30, 1998; revised manuscript received September 10,1998; accepted September 11, 1998.

Address correspondence and reprint requests to Dr. A. C. Nairn atLaboratory of Molecular and Cellular Neuroscience, Rockefeller Uni-versity, New York, NY 10021, U.S.A.

Abbreviations used:Ab, amyloid b protein; AD, Alzheimer’s dis-ease; AEBSF, 4-(2-aminoethyl)benzenesulfonyl fluoride; APP, Alzhei-mer amyloid precursor protein; APPc, APP cytoplasmic domain;CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfon-ate; DTT, dithiothreitol; GST, glutathioneS-transferase; HEK293, hu-man embryonic kidney 293; PBS, phosphate-buffered saline; PI/PTB,phosphotyrosine interaction/phosphotyrosine binding; sAPP, solubleAPP fragment; SDS-PAGE, sodium dodecyl sulfate–polyacrylamidegel electrophoresis; UV-DDB, UV-damaged DNA binding protein.

549

Journal of NeurochemistryLippincott Williams & Wilkins, Philadelphia© 1999 International Society for Neurochemistry

Unreacted sites were blocked with 1M Tris-HCl (pH 7.4) plus50 mM NaCNBH3 for 30 min at room temperature. An addi-tional 8 ml of AminoLink Coupling Gel was also blocked with1 M Tris-HCl (pH 7.4) plus 50 mM NaCNBH3. Six millilitersof treated beads was used as the “preclear” column, and 2 mlwas used as the “control” column. For each purification, brainsfrom six adult rats were homogenized in binding buffer [20 mMHEPES (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA,1 mM dithiothreitol (DTT), 10% glycerin, 1 mM 4-(2-ami-noethyl)benzenesulfonyl fluoride (AEBSF), 5mg/ml chymosta-tin, 5 mg/ml leupeptin, and 5mg/ml pepstatin A] using aPotter–Elvehjem tissue grinder, followed by centrifugation for1 h at 100,000g at 4°C. The supernatant was applied first to the“preclear” column equilibrated with binding buffer, and equalaliquots of the flow-through were applied subsequently to ei-ther the APPc affinity column or to the control column. Afterwashing with 10 ml of binding buffer, elution was performedbatchwise with (a) 0.5M NaCl buffer (binding buffer with 0.5M NaCl), (b) 2M NaCl (binding buffer with 2M NaCl), and (c)6 M guanidine buffer [20 mM HEPES (pH 7.4), 6M guanidine-HCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT). Each elutedfraction was dialyzed against 10 mM HEPES (pH 7.4), 0.1 mMEDTA, 0.1 mM EGTA, 0.1 mM DTT and analyzed by sodiumdodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), followed by silver staining using Rapid-Ag-Stain(ICN).

Protein sequencingEluted proteins were separated by SDS-PAGE and trans-

ferred electrophoretically to Immobilon-P membrane (Milli-pore) as described (Towbin et al., 1979). After the membranewas stained with Ponceau S, a visible protein band at;130kDa was excised and subjected to protein sequencing (Rock-efeller University Protein/DNA Technology Center). The pro-tein was digested with trypsin, and the eluted tryptic peptideswere separated by HPLC as described (Fernandez et al., 1994).Three tryptic peptides were sequenced as described (Athertonet al., 1993).

Preparation of antibodiesAn antibody (CC104) was prepared against a synthetic pep-

tide encompassing residues 1,067–1,080 of monkey UV-DDB/p127 (NH2-KIEHSFWRSFHTER-COOH). Purified peptidewas conjugated to bovine thyroglobulin using glutaraldehyde,and antiserum was prepared in New Zealand White rabbits byCocalico Biologicals Inc. Antibody was affinity-purified byusing a column of synthetic peptide coupled to activated CH-Sepharose. Affinity-purified antibody was concentrated by us-ing a Centriprep Concentrator (Amicon).

Other antibodies used in the study were as follows: G530,raised against residues 597–612 in rat APP695, that recognizesthe extracellular domain of APP (Oishi et al., 1997); and G369,raised against residues 645–694 in APP695, that recognizesAPPc (Buxbaum et al., 1990).

ImmunoblottingProteins were separated by SDS-PAGE and transferred elec-

trophoretically to Immobilon-P membrane as described (Tow-bin et al., 1979). Membranes were incubated in phosphate-buffered saline (PBS) containing 0.1% Tween 20 (Bio-Rad),0.5% nonfat dry milk (Carnation), and antibody. Antibodyconcentrations were;1 mg/ml for G530 and G369 and;5mg/ml for CC104. Binding of the primary antibodies wasdetected by incubation with horseradish peroxidase-conjugatedprotein A (1:2,000–3,000 dilution; Bio-Rad) and the ECL

western blotting detection system (Amersham). Chemilumines-cence was detected by autoradiography. For binding experi-ments using glutathioneS-transferase (GST)-fusion proteinsand cell fractions,125I-labeled protein A (Amersham) was usedfor detection of primary antibody, and radioactivity was quan-tified using a PhosphorImager (Molecular Dynamics).

Co-immunoprecipitationPC12 cells were lysed {;4 3 107 cells/ml of lysis buffer

[PBS plus 10 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 2 mM AEBSF, 5mg/ml chymo-statin, 5mg/ml leupeptin, 5mg/ml pepstatin A, 1 mM Na3VO4,1 mM NaF, and 5mM microcystin]} for 1 h on ice andcentrifuged at 16,000g for 5 min at 4°C. For preclear, normalrabbit IgG (Pierce) was added to the supernatants at a concen-tration of 0.5 mg/ml and incubated on ice for 30 min. The IgGfraction was precipitated by incubation with 100ml of proteinA–Sepharose CL-4B beads (Pharmacia) per milliliter of lysatefor 1 h at4°C, followed by centrifugation. Residual IgG in thesupernatant was precipitated by additional incubation with 100ml of protein A–Sepharose CL-4B beads per milliliter of lysate.For immunoprecipitation of the UV-DDB/p127:APP complex,anti-UV-DDB/p127 antibody (CC104), or antibody that recog-nizes the extracellular domain of APP (G530), was added to theprecleared lysates and incubated on ice for 1 h. ProteinA–Sepharose CL-4B beads were then added and incubated for1 h at 4°C and the samples were subjected to centrifugation.The pellets were washed with lysis buffer three times, andproteins were eluted by boiling the pellet in Laemmli samplebuffer (Laemmli, 1970). Proteins were then analyzed by SDS-PAGE and immunoblotting.

Preparation of GST-APPc fusion proteinsTwo oligonucleotides corresponding to the sense strand, and

three oligonucleotides corresponding to the antisense strand, ofthe DNA encoding residues 649–695 in APP695 were synthe-sized. These sequential oligonucleotides were annealed andligated, then inserted downstream of the thrombin cleavage siteof the pGEX-KT vector [GST-APPc fusion protein, “695(wt)”in Fig. 3A] (Hakes and Dixon, 1992). DNA fragments forGST-truncated APPc fusion proteins were prepared by PCRusing the above plasmid as a template and inserted into thepGEX-KT vector. Point mutations within APPc of the aboveGST-APPc fusion protein were generated using a QuickChangeSite-Directed Mutagenesis Kit (Stratagene). Production andpurification of the GST-fusion proteins were carried out asdescribed (Hakes and Dixon, 1992). Glutathione Sepharose 4Bbeads (Pharmacia) bearing the GST-fusion proteins werewashed with PBS three times. To quantify the bound fusionproteins, a small aliquot of the beads bearing each of the fusionproteins was incubated with PBS containing 10 mM reducedglutathione for 20 min at room temperature, and the concen-tration of eluted fusion protein was measured by absorbance at280 nm. The amounts of the different fusion proteins wereadjusted to a constant level of protein per bed volume byaddition of unbound glutathione Sepharose 4B beads.

Preparation of cytosolic and nuclear fractions fromPC12 cells

Cytosolic and nuclear fractions were prepared from PC12cells as described (Dignam, 1990). Whole-cell extracts wereprepared by lysing the cells with 2% SDS, 62.5 mM Tris-HCl(pH 6.8). Protein concentration was determined using bicincho-ninic acid protein assay reagents (Pierce) with bovine serumalbumin as a standard.

J. Neurochem., Vol. 72, No. 2, 1999

550 T. WATANABE ET AL.

In vitro binding of UV-DDB/p127 to GST-APPcfusion proteins

Glutathione Sepharose beads bearing various GST-fusionproteins were incubated for 2 h at 4°C with the cytosolicfraction of PC12 cells diluted to;400 mg/ml protein in PBSplus 10 mM CHAPS, 2 mM AEBSF, 5mg/ml chymostatin, 5mg/ml leupeptin, and 5mg/ml pepstatin A. As a negativecontrol, glutathione Sepharose beads bearing GST were incu-bated with the cytoplasmic fraction of PC12 cells. The beadswere washed with binding buffer three times, and proteins wereeluted with PBS containing 10 mM reduced glutathione. Fol-lowing SDS-PAGE, the proteins were analyzed by immuno-blotting with 125I-labeled protein A and quantified using aPhosphorImager. In some samples, APPc peptide (residues645–694 of APP695, 100 mM) was added for 1 h before theaddition of the glutathione beads. Binding activities of theGST-fusion proteins were corrected for background determinedwith GST alone. Experiments were done at least in quadrupli-cate, and statistical analyses were performed by ANOVA withScheffe’s F procedure.

ImmunocytochemistryHuman embryonic kidney 293 (HEK293) cells were fixed

and stained with anti-UV-DDB/p127 antibody (CC104) andobserved under a confocal laser scanning microscope as de-scribed (Tomita et al., 1998). Rat hippocampal neurons inculture were prepared from embryonic day 18 rats as described(Bartlett and Banker, 1984). Differentiated neurons were fixedfor 10 min at room temperature with 4% (wt/vol) paraformal-dehyde in PBS containing 4% (wt/vol) sucrose and permeabil-ized for 5 min at room temperature with 0.2% (vol/vol) TritonX-100 in PBS. The cells were incubated with primary antibod-ies [anti-UV-DDB/p127 antibody (CC104) and anti-a-tubulinantibody (TU-01; Zymed)] for 12 h at 4°C, washed well, andthen incubated for 1 h at room temperature with rhodamine-conjugated (for UV-DDB/p127) or fluorescein isothiocyanate-conjugated (fora-tubulin) secondary antibodies.

RESULTS

Identification of UV-DDB/p127 as an APPc bindingprotein

To identify proteins that interact with APPc, an affin-ity column was prepared by immobilization of a syn-thetic peptide encompassing APPc (residues 645–694 ofAPP695). AminoLink coupling gel without immobilizedpeptide was used as a control column. To minimizecontamination of proteins that nonspecifically bind to thegel matrix, the cytosolic fraction of the rat brain waspassed through a preclear column prepared in the sameway as the control column. Proteins were eluted fromeither the affinity or the control column with 6M gua-nidine and analyzed by SDS-PAGE, followed by silverstaining. A protein of;130 kDa was identified only inthe APPc affinity column eluate (Fig. 1A, lane 2). Noother protein was specifically eluted from the APPcaffinity column. A protein of;50 kDa that was detectedin both the APPc affinity column eluate and the controlcolumn eluate (Fig. 1A, lanes 1 and 2) was identified astubulin by protein sequencing and immunoblotting (datanot shown).

The;130-kDa protein eluted with 6M guanidine wassubjected to amino acid sequencing. Three tryptic frag-

ments were isolated by HPLC and sequenced. A searchof the Genbank database showed that the three partialamino acid sequences, LVFSNVNLK, PNTYFIVG-TAMVYPE, and IEHSFWRSFHTERK, were identicalto residues 666–674, 825–839, and 1,068–1,081, re-spectively, of UV-DDB cloned from monkey CV-1 cells(Takao et al., 1993). The calculated molecular mass ofUV-DDB, 127 kDa, was in good agreement with;130kDa, the apparent molecular mass of the protein identi-fied by APPc affinity column chromatography. We con-clude that the protein identified by APPc affinity columnchromatography is a rat homologue of UV-DDB. Basedon co-purification studies, UV-DDB was found to becomposed of a 127-kDa subunit and a 48-kDa subunit(reported to show an apparent molecular mass of 41 kDausing SDS-PAGE) (Keeney et al., 1993; Dualan et al.,1995). However, we failed to identify a 41-kDa proteinin APPc affinity column eluates (Fig. 1A, lane 2).

Antibody against UV-DDB/p127 (CC104) was pre-pared by immunization of rabbits with a synthetic pep-tide that was identical to one of the internal sequencesobtained by protein sequencing of rat UV-DDB/p127(residues 1,067–1,080 of monkey UV-DDB/p127). Pro-teins eluted from the control or APPc affinity columnswere subjected to immunoblotting using CC104. In frac-tions eluted with 6M guanidine, UV-DDB/p127 wasdetected in the eluate from the APPc affinity column, butnot in that from the control column (data not shown). Tocharacterize further the interaction of UV-DDB/p127

FIG. 1. Identification of a protein that interacts with APPc usingaffinity column chromatography. A: Eluates from control andaffinity columns were separated by SDS-PAGE and analyzed bysilver stain. Precleared rat brain cytosol was loaded on a controlcolumn (beads without peptide; lane 1) or an APPc affinity col-umn (beads cross-linked with the APPc peptide; lane 2). Afterextensive washing with binding buffer with 2 M NaCl, proteinswere eluted with 6 M guanidine buffer. A protein of ;130 kDawas eluted only from the APPc affinity column. B: Precleared ratbrain cytosol was loaded on an APPc affinity column. Proteinswere eluted with binding buffer with 0.5 M NaCl (lane 1) or 6 Mguanidine buffer (lane 2) and analyzed by immunoblotting usingthe anti-UV-DDB/p127 antibody. Arrowheads indicate the posi-tion of UV-DDB/p127.


551UV-DDB BINDS TO THE CYTOPLASMIC DOMAIN OF APP

with APP, we analyzed the buffer conditions necessaryto elute UV-DDB/p127 from the APPc affinity column.A significant proportion of bound UV-DDB/p127 waseluted with 0.5M NaCl buffer, with the remaining boundUV-DDB/p127 being eluted with 6M guanidine (Fig.1B).

Immunoprecipitation of UV-DDB/p127:APP fromPC12 cells

To examine whether APP and UV-DDB/p127 inter-acted in cells, co-immunoprecipitation experiments wereperformed. PC12 cells were lysed in the presence of 10mM CHAPS. The detergent-soluble fraction was pre-cleared with normal rabbit IgG and protein A–Sepharoseand subjected to immunoprecipitation using affinity-pu-rified anti-APP extracellular domain antibody (G530) oranti-UV-DDB/p127 antibody (CC104). The same quan-tity of normal rabbit IgG was used as a control. Theimmunoprecipitates were analyzed by immunoblottingusing CC104 or anti-APPc antibody (G369) (Fig. 2).UV-DDB/p127 was detected in the sample immunopre-cipitated with G530 (Fig. 2, lane 4). Both mature andimmature forms of APP were detected in the sampleimmunoprecipitated with CC104 (Fig. 2, lane 2). Theratio of mature APP and immature APP co-immunopre-cipitated with UV-DDB/p127 was comparable to that inthe cell lysate used as a starting material (Fig. 2, lane 1).UV-DDB/p127 or APP could not be detected in thesample immunoprecipitated using normal rabbit IgG,indicating that the co-immunoprecipitation of UV-DDB/p127 and APP was specific to the respective antibodies(Fig. 2, lanes 3 and 5). The PC12 cells used in these

experiments were normal untransfected cells. Therefore,UV-DDB/p127 and APP interact in lysates from normalmammalian cells, raising the possibility that they interactunder physiological conditions in mammalian cells. TheUV-DDB/p127:APP complex was also co-immunopre-cipitated from cell lysates of HEK293 cells overexpress-ing human APP770. Notably, in cells expressing a form ofhuman APP770 truncated at residue 726 (which corre-sponds to residue 651 in APP695), the amount of UV-DDB/p127 detected in immunoprecipitates was reducedto background levels (data not shown).

Characterization of the interaction betweenUV-DDB/p127 and a GST-APPc fusion protein

A GST-APPc fusion protein (residues 649–695 inAPP695) and several deletion mutants (shown in Fig. 3A)were individually immobilized to glutathione Sepharosebeads, and the beads were incubated with a cytosolicfraction from PC12 cells. GST was immobilized to glu-tathione Sepharose beads and used as a negative control.UV-DDB/p127 bound to the GST-APPc fusion protein,and this interaction was abolished by the addition of apeptide encompassing residues 645–694 of APP695 (Fig.3B). Analysis of the various mutants revealed that dele-tion of 14 or more amino acids at the C terminus of APPsignificantly reduced the interaction. Notably, the final16 amino acids of APP showed partial binding activity,supporting the idea that the C-terminal region of APP iscritical to the interaction with UV-DDB/p127. However,this result indicates that amino acids 649–679 contributeeither directly or indirectly to the interaction of UV-DDB/p127 with APP. The C-terminal region of APPcontains the motif, YENPTY, that has been shown to beimportant for endocytosis of the protein (Koo andSquazzo, 1994; Lai, et al., 1995). To examine the con-tribution of the YENPTY motif to the interaction withUV-DDB/p127, single amino acid substitutions withinthis sequence were introduced into the GST-APPc fusionprotein. Substitution of Tyr682 with glycine (Y682G)inhibited the interaction (Fig. 3B). Substitution of Asp684

(N684A), or Tyr687 (Y687A), with alanine inhibited theinteraction to a lesser extent than that of the Y682Gmutation.

Intracellular localization of UV-DDB/p127Although UV-DDB/p127 has been isolated previously

from nuclear extracts (Abramic et al., 1991), the intra-cellular localization has not been analyzed in detail. Toexamine the intracellular localization of UV-DDB/p127,cytosolic and nuclear fractions were prepared and ana-lyzed by immunoblotting with anti-UV-DDB/p127 anti-body (Fig. 4A). Most of the UV-DDB/p127 was recov-ered in the cytosolic fraction (Fig. 4A, lane 2), althoughsome UV-DDB/p127 was present in the nuclear extract(Fig. 4A, lane 3). The presence of UV-DDB/p127 in thecytosolic fraction could not be explained by contamina-tion with nucleoplasm during fractionation, because im-munoblotting of the same fractions with antibody againstU-70 splicing factor indicated that this polypeptide was

FIG. 2. Co-immunoprecipitation of UV-DDB/p127 and APP fromPC12 cell lysate. PC12 cells were lysed in PBS plus 10 mMCHAPS, and the lysate (lane 1) was used for immunoprecipita-tion with anti-UV-DDB/p127 antibody (ap127, CC104) (lane 2),anti-APP extracellular domain antibody (aAPP, G530) (lane 4), ornormal rabbit IgG (lanes 3 and 5). The immunoprecipitates wereanalyzed by immunoblotting with anti-APPc antibody (aAPP,G369) (lanes 1–3) or anti-UV-DDB/p127 antibody (ap127,CC104) (lanes 4 and 5). Mature APP (mAPP) and immature APP(imAPP) were co-immunoprecipitated with UV-DDB/p127 usingthe CC104 antibody (lane 2, filled arrowheads). The designationof immature and mature isoforms of APP was based on theoriginal description by Weidemann et al. (1989) and also on ourprevious studies (Caporaso et al., 1992). The ratio betweenmature APP and immature APP in the co-precipitate was com-parable to that in the cell lysate (lane 1). UV-DDB/p127 was alsoco-immunoprecipitated with APP using the G530 antibody (lane4, open arrowhead). IP, immunoprecipitation.



detected only in the nuclear fraction (data not shown). Acytoplasmic localization of UV-DDB/p127 was also re-vealed by immunocytochemical studies (Fig. 4B–D).

DISCUSSION

We have identified, by affinity column chromatogra-phy, a;130-kDa protein from rat brain cytosol, whichinteracts with APPc. Amino acid sequencing of the;130-kDa protein showed it to be a rat homologue ofUV-DDB/p127. Furthermore, our studies have indicatedthat both anti-APP antibody and anti-UV-DDB/p127 an-tibodies immunoprecipitated the UV-DDB/p127:APPcomplex from cell lysates. These results suggest thatUV-DDB/p127, which is localized primarily to the cy-tosolic fraction, interacts with APPc. UV-DDB was iden-tified originally as a factor that binds to UV-damagedDNA (Hirschfeld et al., 1990; Abramic et al., 1991;Takao et al., 1993). Some studies have reported thatUV-DDB is a heterodimer of subunits of 127 and 48 kDa(Keeney et al., 1993; Dualan et al., 1995). However,consistent with a more recent study (Aboussekhra et al.,1995), no such polypeptide was identified in eluates fromthe APPc affinity column.

In addition to UV-DDB/p127, several proteins havebeen characterized recently as APPc binding proteins.Interaction of the Go protein with APP was revealed by

co-immunoprecipitation studies (Nishimoto et al., 1993),and it has been suggested that the APP:Go interactionmay be involved in apoptosis (Yamatsuji et al., 1996).APP-BP1, which was cloned by screening a cDNA ex-pression library with an APPc peptide, is homologous toAXR1, a protein believed to be involved in signal trans-duction in plant cells (Chow et al., 1996). Yeast two-hybrid analyses have identified rat Fe65 (Fiore et al.,1995) and its human homologues (Gue´nette et al., 1996)as APP binding proteins. In Fe65, two phosphotyrosineinteraction/phosphotyrosine binding (PI/PTB) domainsappear to mediate the interaction with the C-terminalregion of APP (Zambrano et al., 1997). X11a and b,other proteins with a PI/PTB domain, have also beenshown to interact with APPc (Borg et al., 1996, 1998;McLoughlin and Miller, 1996). These findings suggestthat APP is involved in signal transduction, possibly as acell surface receptor.

In addition to, or in a manner distinct from, theirpotential involvement in signal transduction, APPc bind-ing proteins may affect APP metabolism. For example,the various APP binding proteins could affect traffickingof APP from the endoplasmic reticulum to the Golgi,from the Golgi to the plasma membrane, or through theendocytic pathway. In this respect, the PI/PTB domain(s)of Fe65 and X11 have been shown to interact with the

FIG. 3. Binding of UV-DDB/p127 to APPc. A: Vari-ous truncations and point mutations were intro-duced into a GST-APPc wild-type fusion protein[residues 649–695 of APP695; 695(wt)]. Substitutedamino acid residues are underlined. B: Glutathionebeads bearing GST-fusion proteins were incubatedwith the cytosolic fraction of PC12 cells, and UV-DDB/p127 bound was measured by immunoblottingusing anti-UV-DDB/p127 antibody. APPc peptide[residues 645–694 of APP695; 695(wt) 1 peptide]was added to the cytosolic fraction at a concentra-tion of 100 mM. The binding activity of each GST-fusion protein is expressed as a percentage of thebinding activity for the GST-APPc wild-type fusionprotein [695(wt)]. Error bars indicate the SEM values[for 695(wt) and GST, n 5 12; for Y682G, N684A, andY687A, n 5 8; for 681s, 671s, 661s, 651s, 680–695,and 695(wt) 1 peptide, n 5 4], and asterisks indicatesignificant differences (**p , 0.001; ***p , 0.0001)from 695(wt) used as a control. The value obtainedfor 680–695 was significantly different from that forGST ( p , 0.02). The value obtained for Y682G wasnot significantly different from that for GST.



YENPTY motif in APPc. Deletion of the YENPTY motifresults in increased secretion of the soluble APP frag-ment (sAPP) while decreasing the production of Ab(Koo and Squazzo, 1994; Lai et al., 1995). It has beensuggested that a pool of Ab is generated within theendocytic pathway and that the YENPTY motif is im-portant for the internalization of APP leading to Absecretion. Therefore, Fe65 and X11 may be involved inthe internalization of APP and/or production of Ab in theendocytic pathway. However, it has been reported re-cently that the overexpression of X11a or b slowedcellular APP metabolism and reduced production of bothsAPP and Ab (Borg et al., 1998), suggesting that theconsequences of interaction of APP with its different

binding proteins may be complex. The present studiesindicated that deletion of the C-terminal 14 amino acidsencompassing the YENPTY motif significantly reducedthe interaction of APP with UV-DDB/p127. Further-more, mutation of Tyr682 (Y682G) inhibited binding ofAPP to UV-DDB/p127. In contrast, mutation of Asp684

(N684A) or Tyr687 (Y687A) had less effect than that ofmutation of Y682G. These results are similar to thatobtained for Fe65 (Borg et al., 1996), although there isno obvious homology between UV-DDB/p127 and thePI/PTB domains of Fe65. We observed that the overex-pression of UV-DDB/p127 slightly increased the secre-tion of sAPP and that the overexpression of its antisensemRNA slightly decreased secretion of sAPP (S. Tomita

FIG. 4. Intracellular localization ofUV-DDB/p127. A: Subcellularfractions of PC12 cells were pre-pared as described in Experimen-tal Procedures. Aliquots of frac-tions, representing similar num-bers of cells, were analyzed byimmunoblotting using anti-UV-DDB/p127 antibody (CC104).Lane 1, whole cell lysate; lane 2,cytosolic fraction; and lane 3, nu-clear extract. UV-DDB/p127 wasdetected mainly in the cytosolicfraction (lane 2), with a smallamount being detected in the nu-clear extract (lane 3). The arrow-head indicates the position of UV-DDB/p127. B: HEK293 cells werestained with anti-UV-DDB/p127antibody (CC104) and observedunder a confocal laser scanningmicroscope. C and D: Primarycultures of hippocampal neuronswere double-stained for UV-DDB/p127 (C) and a-tubulin (D).



and T. Suzuki, unpublished observations). These resultsmight be explained by mutually exclusive binding ofUV-DDB/p127 or other APP binding proteins that slowAPP cellular processing, such as X11, to the YENPTYmotif of APP.

It has been proposed that UV-DDB is involved innuclear excision repair, a feature of DNA damage by UVirradiation, based on the following observations: (a) thebinding of UV-DDB to UV-damaged DNA has beenreported to be absent in fibroblasts from certain patientsof the xeroderma pigmentosum group E (XP-E) (Chuand Chang, 1988), in which deficits in DNA repair leadto a congenital skin disease; and (b) microinjection ofpurified UV-DDB into cell lines lacking UV-DDB ac-tivity (XP-E[DDB2] cells) was reported to restore theDNA repair defect (Keeney et al., 1994). However, re-cent in vitro studies of reconstituted nuclear excisionrepair have revealed that UV-DDB is not an essentialfactor for the core processes of the repair system(Aboussekhra et al., 1995; Mu et al., 1995; Kazantsev etal., 1996). Based on these recent studies, it has beensuggested that UV-DDB may play a role as an accessoryfactor that facilitates the core processes of DNA repair.For example, it may function as a molecular chaperone tofacilitate assembly of the DNA repair machinery (Ka-zantsev et al., 1996).

In certain patients with congenital deficiencies inDNA repair, i.e., xeroderma pigmentosum, ataxia telan-giectasia, and Cockayne syndrome, progressive neuronaldegeneration is observed (Mazzarello et al., 1992). Con-sidering the fact that postmitotic neurons do not replicateand have relatively low levels of DNA repair enzymes, itwas suggested that a deficit in DNA repair may renderthe nervous system susceptible to DNA damage, leadingto neuronal degeneration (Mazzarello et al., 1992). No-tably, lower DNA repair activity in cells from familialand sporadic AD patients compared with age-matchedcontrols has been reported (Jones et al., 1989; Boerrigteret al., 1992; Parshad et al., 1996). Although this may notbe a primary cause of AD, it still could be involved in theprocess of neuronal degeneration in AD. In this regard,malfunction of UV-DDB/p127 could enhance the accu-mulation of DNA damage and, in turn, contribute toneuronal degeneration.

Recent studies of UV-DDB/p127 have identified po-tential functions for the protein other than a role in DNArepair. UV-DDB/p127 is highly related to the transcrip-tion activator, apolipoprotein B gene regulatory factor-2(Krishnamoorthy et al., 1997), a protein that binds to aregulatorycis-element of the apolipoprotein B100 gene(Zhuang et al., 1992). The ubiquitous expression of UV-DDB/p127 (Takao et al., 1993) suggests that if thisprotein also acts as a transcription factor, its actions areunlikely to be restricted to the regulation of apolipopro-tein B100 gene expression, which occurs primarily in theliver (Das et al., 1988). Furthermore, it has been reportedthat a complex of UV-DDB/p127 and the p48 subunitinteracts with the transcription factor, E2F1, via the p48subunit (Hayes et al., 1998). Our present studies have

indicated that UV-DDB/p127 is localized to both thecytosol and nucleus. Possibly, UV-DDB/p127 may trans-locate between the cytosol and nucleus as part of a signaltransduction process involved in the regulation of genetranscription. Future studies will be necessary to eluci-date the precise function of UV-DDB/p127 and to un-derstand its potential role in APP biology.

Acknowledgment: We thank Gloria Bertuzzi for technicalassistance. Protein sequence data were obtained at the Rock-efeller University Protein/DNA Technology Center. This workwas supported by USPHS grant AD 09464 (A.C.N. and P.G.).This work was supported in part by a grant from the Programfor Promotion of Basic Research Activities for InnovativeBiosciences Japan (T.S.).

REFERENCES

Aboussekhra A., Biggerstaff M., Shivji M. K., Vilpo J. A., MoncollinV., Podust V. N., Protic M., Hubscher U., Egly J. M., and WoodR. D. (1995) Mammalian DNA nucleotide excision repair recon-stituted with purified protein components.Cell 80, 859–868.

Abramic M., Levine A. S., and Protic M. (1991) Purification of anultraviolet-inducible, damage-specific DNA-binding protein fromprimate cells.J. Biol. Chem.266,22493–22500.

Atherton D., Fernandez J., DeMott M., Andrews L., and Mische S.(1993) Routine protein sequence analysis below ten picomoles:one sequencing facilities approach, inTechniques in ProteinChemistry IV(Angeletti R. H., ed), pp. 409–418. Academic Press,San Diego.

Bartlett W. P. and Banker G. A. (1984) An electron microscopic studyof the development of axons and dendrites by hippocampal neu-rons in culture. I. Cells which develop without intercellular con-tacts.J. Neurosci.4, 1944–1953.

Boerrigter M. E., Wei J. Y., and Vijg J. (1992) DNA repair andAlzheimer’s disease.J. Gerontol.47, B177–B184.

Borg J. P., Ooi J., Levy E., and Margolis B. (1996) The phosphoty-rosine interaction domains of X11 and FE65 bind to distinct siteson the YENPTY motif of amyloid precursor protein.Mol. Cell.Biol. 16, 6229–6241.

Borg J. P., Yang Y., Tadde´o-Borg M., Margolis B., and Turner R. S.(1998) The X11a protein slows cellular amyloid precursor proteinprocessing and reduces Ab40 and Ab42 secretion.J. Biol. Chem.273,14761–14766.

Buxbaum J. D., Gandy S. E., Cicchetti P., Ehrlich M. E., Czernik A. J.,Fracasso R. P., Ramabhadran T. V., Unterbeck A. J., and Green-gard P. (1990) Processing of Alzheimer beta/A4 amyloid precur-sor protein: modulation by agents that regulate protein phosphor-ylation. Proc. Natl. Acad. Sci. USA87, 6003–6006.

Caporaso G. L., Gandy S. E., Buxbaum J. D., Ramabhadran T. V., andGreengard P. (1992) Protein phosphorylation regulates secretionof Alzheimer beta/A4 amyloid precursor protein.Proc. Natl.Acad. Sci. USA89, 3055–3059.

Chow N., Korenberg J. R., Chen X. N., and Neve R. L. (1996)APP-BP1, a novel protein that binds to the carboxyl-terminalregion of the amyloid precursor protein.J. Biol. Chem.271,11339–11346.

Chu G. and Chang E. (1988) Xeroderma pigmentosum group E cellslack a nuclear factor that binds to damaged DNA.Science242,564–567.

Das H. K., Leff T., and Breslow J. L. (1988) Cell type-specificexpression of the human apoB gene is controlled by twocis-actingregulatory regions.J. Biol. Chem.263,11452–11458.

Dignam J. D. (1990) Preparation of extracts from higher eukaryotes.Methods Enzymol.182,194–203.

Dualan R., Brody T., Keeney S., Nichols A. F., Admon A., and Linn S.(1995) Chromosomal localization and cDNA cloning of the genes(DDB1 and DDB2) for the p127 and p48 subunits of a humandamage-specific DNA binding protein.Genomics29, 62–69.



Fernandez J., Andrews L., and Mische S. M. (1994) An improvedprocedure for enzymatic digestion of polyvinylidene difluoride-bound proteins for internal sequence analysis.Anal. Biochem.218,112–117.

Fiore F., Zambrano N., Minopoli G., Donini V., Duilio A., and RussoT. (1995) The regions of the Fe65 protein homologous to thephosphotyrosine interaction/phosphotyrosine binding domain ofShc bind the intracellular domain of the Alzheimer’s amyloidprecursor protein.J. Biol. Chem.270,30853–30856.

Guenette S. Y., Chen J., Jondro P. D., and Tanzi R. E. (1996) Asso-ciation of a novel human FE65-like protein with the cytoplasmicdomain of theb-amyloid precursor protein.Proc. Natl. Acad. Sci.USA93, 10832–10837.

Hakes D. J. and Dixon J. E. (1992) New vectors for high levelexpression of recombinant proteins in bacteria.Anal. Biochem.202,293–298.

Hayes S., Shiyanov P., Chen X., and Raychaudhuri P. (1998) DDB, aputative DNA repair protein, can function as a transcriptionalpartner of E2F1.Mol. Cell. Biol. 18, 240–249.

Hirschfeld S., Levine A. S., Ozato K., and Protic M. (1990) A consti-tutive damage-specific DNA-binding protein is synthesized athigher levels in UV-irradiated primate cells.Mol. Cell. Biol. 10,2041–2048.

Jones S. K., Nee L. E., Sweet L., Polinsky R. J., Bartlett J. D., BradleyW. G., and Robison S. H. (1989) Decreased DNA repair infamilial Alzheimer’s disease.Mutat. Res.219,247–255.

Kazantsev A., Mu D., Nichols A. F., Zhao X., Linn S., and Sancar A.(1996) Functional complementation of xeroderma pigmentosumcomplementation group E by replication protein A in an in vitrosystem.Proc. Natl. Acad. Sci. USA93, 5014–5018.

Keeney S., Chang G. J., and Linn S. (1993) Characterization of ahuman DNA damage binding protein implicated in xerodermapigmentosum E.J. Biol. Chem.268,21293–21300.

Keeney S., Eker A. P., Brody T., Vermeulen W., Bootsma D., Hoeij-makers J. H., and Linn S. (1994) Correction of the DNA repairdefect in xeroderma pigmentosum group E by injection of a DNAdamage-binding protein.Proc. Natl. Acad. Sci. USA91, 4053–4056.

Koo E. H. and Squazzo S. L. (1994) Evidence that production andrelease of amyloid beta-protein involves the endocytic pathway.J. Biol. Chem.269,17386–17389.

Krishnamoorthy R. R., Lee T. H., Butel J. S., and Das H. K. (1997)Apolipoprotein B gene regulatory factor-2 (BRF-2) is structurallyand immunologically highly related to hepatitis B virus X asso-ciated protein-1 (XAP-1).Biochemistry36, 960–969.

Laemmli U. K. (1970) Cleavage of structural proteins during theassembly of the head of bacteriophage T4.Nature227,680–685.

Lai A., Sisodia S. S., and Trowbridge I. S. (1995) Characterization ofsorting signals in the beta-amyloid precursor protein cytoplasmicdomain.J. Biol. Chem.270,3565–3573.

Mazzarello P., Poloni M., Spadari S., and Focher F. (1992) DNA repairmechanisms in neurological diseases: facts and hypotheses.J. Neurol. Sci.112,4–14.

McLoughlin D. M. and Miller C. C. (1996) The intracellular cytoplas-mic domain of the Alzheimer’s disease amyloid precursor proteininteracts with phosphotyrosine-binding domain proteins in theyeast two-hybrid system.FEBS Lett.397,197–200.

Mu D., Park C.-H., Matsunaga T., Hsu D. S., Reardon J. T., andSancar A. (1995) Reconstitution of human DNA repair excision

nuclease in a highly defined system.J. Biol. Chem.270,2415–2418.

Neve R. L. (1996) Mixed signals in Alzheimer’s disease.TrendsNeurosci.19, 371–372.

Nishimoto I., Okamoto T., Matsuura Y., Takahashi S., Okamoto T.,Murayama Y., and Ogata E. (1993) Alzheimer amyloid proteinprecursor complexes with brain GTP-binding protein G(o).Nature362,75–79.

Oishi M., Nairn A. C., Czernik A. J., Lim G. S., Isohara T., GandyS. E., Greengard P., and Suzuki T. (1997) The cytoplasmic do-main of Alzheimer’s amyloid precursor protein is phosphorylatedat Thr654, Ser655, and Thr668 in adult rat brain and cultured cells.Mol. Med.3, 111–123.

Parshad R. P., Sanford K. K., Price F. M., Melnick L. K., Nee L. E.,Schapiro M. B., Tarone R. E., and Robbins J. H. (1996) Fluores-cent light-induced chromatid breaks distinguish Alzheimer diseasecells from normal cells in tissue culture.Proc. Natl. Acad. Sci.USA93, 5146–5150.

Selkoe D. J. (1996) Amyloid beta-protein and the genetics of Alzhei-mer’s disease.J. Biol. Chem.271,18295–18298.

Takao M., Abramic M., Moos M. Jr., Otrin V. R., Wootton J. C.,McLenigan M., Levine A. S., and Protic M. (1993) A 127 kDacomponent of a UV-damaged DNA-binding complex, which isdefective in some xeroderma pigmentosum group E patients, ishomologous to a slime mold protein.Nucleic Acids Res.21,4111–4118.

Tomita S., Kirino Y., and Suzuki T. (1998) Cleavage of Alzheimer’samyloid precursor protein (APP) by secretases occurs afterO-glycosylation of APP in the protein secretory pathway. Identifi-cation of intracellular compartments in which APP cleavage oc-curs without using toxic agents that interfere with protein metab-olism. J. Biol. Chem.273,6277–6284.

Towbin H., Staehelin T., and Gordon J. (1979) Electrophoretic transferof proteins from polyacrylamide gels to nitrocellulose sheets.Proc. Natl. Acad. Sci. USA76, 4350–4354.

Weidemann A., Konig G., Bunke D., Fischer P., Salbaum J. M.,Masters C. L., and Beyreuther K. (1989) Identification, biogene-sis, and localization of precursors of Alzheimer’s disease A4amyloid protein.Cell 57, 115–126.

Yamatsuji T., Matsui T., Okamoto T., Komatsuzaki K., Takeda S.,Fukumoto H., Iwatsubo T., Suzuki N., Asami-Odaka A., IrelandS., Kinane T. B., Giambarella U., and Nishimoto I. (1996) Gprotein-mediated neuronal DNA fragmentation induced by famil-ial Alzheimer’s disease-associated mutants of APP.Science272,1349–1352.

Younkin S. G. (1995) Evidence that A beta 42 is the real culprit inAlzheimer’s disease.Ann. Neurol.37, 287–288.

Zambrano N., Buxbaum J. D., Minopoli G., Fiore F., De Candia P., DeRenzis S., Faraonio R., Sabo S., Cheetham J., Sudol M., andRusso T. (1997) Interaction of the phosphotyrosine interaction/phosphotyrosine binding-related domains of Fe65 with wild-typeand mutant Alzheimer’s beta-amyloid precursor proteins.J. Biol.Chem.272,6399–6405.

Zhuang H., Chuang S. S., and Das H. K. (1992) Transcriptionalregulation of the apolipoprotein B100 gene: purification and char-acterization oftrans-acting factor BRF-2.Mol. Cell. Biol. 12,3183–3191.



a 127-kda protein (uv-ddb) binds to the cytoplasmic domain of the alzheimer's amyloid precursor...

Documents