NAPP2, a Peroxisomal Membrane Protein, Is Also a Transcriptional Corepressor

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

Article

NAPP2, a Peroxisomal Membrane Protein, Is Also a Transcriptional Corepressor Narender R. Gavva,5,* Shau-Ching Wen,1 Pratibha Daftari,5 Mariko Moniwa,2 Wen-Ming Yang,3 Lan-Ping Teresa Yang-Feng,4 Edward Seto,3 James R. Davie,2 and Che-Kun James Shen1,5,† 1 Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan, Republic of China Department of Biochemistry and Molecular Biology, University of Manitoba, Manitoba R3E OW3, Canada 3 Moffitt Cancer Center and Research Institute, Department of Medical Microbiology and Immunology, College of Medicine, University of South Florida, Tampa, Florida 33612, USA 4 Department of Medicine, Yale School of Medicine, New Haven, Connecticut 06520-8024, USA 5 Section of Molecular and Cellular Biology, University of California, Davis, California 95616, USA 2

*Present address: Amgen Inc., Thousand Oaks, California 91362, USA †

To whom correspondence and reprint requests should be addressed. Fax: 8862 2788 4177 or (530) 752-3085. E-mail: [email protected] or [email protected].

Nuclear factor-erythroid number 2 (NF-E2) is a positive regulatory, DNA binding transcription factor for gene expression in erythroid and megakaryocytic cells. To further understand the mechanisms of NF-E2 function, we used expression cloning to identify coregulators interacting with the erythroid-specific subunit of NF-E2, p45. We have isolated a protein, NAPP2, which contains an aspartic-acid- and glutamic-acid-rich region and a nuclear localization signal. The gene encoding NAPP2, PEX14, is located on chromosome 1p36 and is ubiquitously expressed. The domains of interaction in vitro and in vivo between p45 and NAPP2 were mapped by a yeast two-hybrid system and cotransfection experiments. In mammalian cell culture, ectopically expressed NAPP2 inhibited p45-directed transcriptional activation. Furthermore, NAPP2 functions as a corepressor and interacts specifically with histone deacetylase l (HDAC1), but not HDAC2 or HDAC3. NAPP2 is thus potentially a negative coregulator of NF-E2. NAPP2 is identical to PEX14, an integral membrane protein essential for protein docking onto the peroxisomes. These studies have identified a novel, bifunctional protein capable of acting as a transcriptional corepressor and a polypeptide transport modulator. They also suggest that NF-E2 may function both positively and negatively in the transcription regulation of specific erythroid and megakaryocytic genes.

INTRODUCTION Nuclear factor-erythroid number 2 (NF-E2) is an obligate heterodimer between the hematopoietic-specific p45 and widely expressed Maf/p18 subunits, both members of the basic region-leucine zipper (bZIP) family of transcription factors [1–3]. All data available thus far have suggested that NF-E2 is a positive regulator of transcription. Using a dominantnegative mutant of p18/NF-E2 to inhibit NF-E2 activity in mouse erythroleukemia (MEL) cells, it has been demonstrated [4] that reduced levels of both ␣- and ␤maj-globins were associated with a lower level of NF-E2 activity. Globin expression was restored upon the introduction of a tethered p45–p18 heterodimer. CB3, a MEL-derived cell line, does not express p45 because of a retroviral integration of the Friend leukemia virus genome in the p45 gene [5]. Unlike other MEL cells, no

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved. 0888-7543/02 $35.00

detectable levels of ␣- and ␤maj-globin mRNA were observed in CB3 cells treated with dimethyl sulfoxide (DMSO). Partial restoration of depressed globins was observed upon reintroduction of p45/NF-E2 cDNA [5]. These data further supported the role of p45/NF-E2 being an activator of transcription in erythroid and megakaryocytic cells. Consistent with the above role of NF-E2, an activation domain of p45 has been identified at amino acids 1–206 [4]. This region is capable of interacting with the histone acetyl transferase and coactivator CBP [6], the TBP-associated factor TAF II 130/110 [7], and a group of WW-domain containing proteins, among which are several transcription coactivators [8]. An in vivo role by NF-E2 to disrupt/remodel local nucleosomal structure at the vicinity of the NF-E2/AP1 motif(s) has also been suggested [9,10]. This process would also facilitate the transcription processes. NF-E2 exerts its

423

Article

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

A

NAPP2

FIG. 1. Map and sequence of NAPP2. (A) Structural features of NAPP2 and p45. The activation domain and basic leucine zipper region of p45, as well as the acidic region (DE) and nuclear localization signal-like sequence NLS, in NAPP2 are indicated. The dotted lines indicate the interacting regions of p45 and NAPP2, respectively. (B) Sequence alignment of NAPP2 cDNA with the translated protein sequence. The NLS is boxed, and the polyadenylation signal AATAAA is underlined.

B NAPP2 is the human homolog of Pex14p, a yeast protein characterized and demonstrated to be required for polypeptide docking on the peroxisomes [18–20].

RESULTS

function through binding to the NF-E2/AP1 sequence motifs with the consensus 5’-GCTGAG/CTCA-3⬘, which exists in different transcription regulatory elements including the locus control region (LCR) of the mammalian ␤-globin loci [11], the HS-40 enhancer of the mammalian ␣-globin loci [12], and promoters of various megakaryocytic genes such as those encoding porphobilinogen deaminase (PDBG) [13] and the thromboxane synthetase [14]. Indeed, NF-E2 is associated with the former two elements in living erythroid cells [15,16]. To better understand the mechanisms of transcriptional activation and chromatin remodeling by p45/NF-E2, we have attempted to isolate proteins that interact with NF-E2. Previously, one such protein has been identified to be human ITCH, a ubiquitin ligase [17]. We now report the use of expression cloning to isolate another gene (PEX14) encoding the protein NAPP2 (NF-E2 associated polypeptide-2). NAPP2 interacts with the activation domain of p45 in vitro and in vivo, and it behaves as a corepressor for p45-mediated transcriptional activation. The repression seems to be modulated through one or both of two different mechanisms. Thus, NFE2 could also function as a negative regulator of transcription through the recruitment of corepressors such as NAPP2.

424

Molecular Cloning of NAPP2 cDNA To identify cDNAs encoding proteins that bind p45, we screened a K562 cDNA library in a ␭-phage expression vector with a radiolabeled recombinant GSTp45/NF-E2 protein. Phage clones that produced proteins capable of binding to the recombinant p45 were then detected by autoradiography. A single positive clone was identified after screening of 106 phage clones. The 1.9-kb cDNA insert from this phage was amplified by PCR using Vent Polymerase (New England Biolabs). DNA sequencing revealed a 1131-bp coding region in-frame with the ␤-galactosidase in the phage (Fig. 1; GenBank acc. no. AB017546). This cDNA and the encoded protein were named NAPP2. NAPP2 seems to be identical to a human peroxisomal membrane protein, PEX14. It also exhibits low homology with a yeast hypothetical ORF, YGL 153W (GenBank acc. no. 272675), and one Caenorhabditis elegans ORF, R07H5.1

FIG. 2. RT-PCR analysis of NAPP2 expression. cDNAs prepared from RNAs of different cell lines were used for RT-PCR analysis. Each PCR reaction contained cDNA from 100 ng of total RNA. As a control, a ␤-actin fragment was co-amplified in the same PCR reaction mixtures. The number of PCR cycles, 30, was chosen to give a linear range of the amplification.

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

A

Article

FIG. 3. p45–NAPP2 interaction in vitro. (A) Far-western analysis of NAPP2–p45 interaction. Purified plaques of ␤-gal-expressing or ␤-gal-NAPP2-expressing ␭gt11 phages were grown in E. coli Y1089 and induced at 42⬚C. Total lysate proteins (100 ␮g) were separated on SDS-PAGE gels and duplicate blotted. One of the blots was probed with 32P-labeled GST–p45 as shown. (B) Anti-␤-gal western blot. The second blot from (A) was stained with anti-␤-gal antibody. (C) Mapping of the NAPP2 domain interacting with p45. (Left) Coomassiestained gel of different GST fusions of NAPP2. (Right) Autoradiograph of the gel blot hybridized with 32P-labeled GST–p45. The 26-kDa bands in lanes 1 and 2 of the left panel are degradation products from the fusion polypeptides. Also, note that both GST–NAPP2 and GST–NAPP2(203–373) migrated more slowly than expected from their molecular weights. The anomalous gel behaviors of NAPP2 have also been noted by others (Fig. 9 of [29]), and are likely due to the acidic nature of the C terminus.

B

C

(GenBank acc. no. Z81107). As determined with the PSORT program, a putative nuclear localization signal consensus, RRGGDGQINEQVEKLRR, is located between amino acids 350 and 366. NAPP2 possesses an aspartic acid (D)- and glutamic acid (E)-rich region near its carboxy-terminal end (Fig. 1). Such D/E-rich regions have been identified in several proteins implicated in chromatin remodeling [21]. The NAPP2 amino acid sequence is identical to that of a recently cloned human peroxisomal membrane-binding protein, PEX14. Expression and Chromosomal Localization of PEX14 We examined the expression patterns of human PEX14 in different cell lines including K562, HeLa, Dami, and KG1. The PEX14 transcript was detected by RT-PCR in all cell types studied (Fig. 2). It seems that NAPP2 is downregulated dur-

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

ing phorbol 12-myristate 13-acetate (PMA) induction of Dami, a derivative of the human erythroleukemia cell line HEL [22], toward the megakaryocytic lineage. It is notable that p45 is known to be necessary for the megakaryocytic differentiation process [23], and that NAPP2 is a repressor of p45-mediated transcriptional activation. Northern blot analysis with 50 ␮g of total RNA from K562 and HeLa cell lines showed a weak transcript of approximately 2000 nt, and genomic Southern blot analysis indicated that NAPP2 is encoded by a single copy gene (data not shown). FISH analysis indicated that PEX14 is localized on chromosome 1p36. p45–NAPP2 Interaction in Vitro To examine the specific association between p45 and NAPP2, we used the lysogen to produce NAPP2 as a fusion protein with ␤-galactosidase (␤-gal–NAPP2). Protein lysates of ␤-gal and ␤-gal–NAPP2 lysogens were separated by SDSPAGE in duplicate and then transferred to nitrocellulose filters. Probing with 32P-labeled GST–p45 detected an 170kDa band in the lanes containing lysate from the ␤-gal lysogen (Fig. 3A, lanes 1, 2, 5, and 6), but not in those from the ␤-gal lysogens (Fig. 3A, lanes 3 and 4). This result was confirmed with the use of the ␤-gal antibody in western analysis (Fig. 3B). The additional bands in lanes 1, 2, 5, and 6 of Fig. 3B most likely represent degraded products of ␤-gal–NAPP2. Using far-western analysis, we have further mapped the region of NAPP2 interacting with p45. Nitrocellulose membrane blotted with GST–NAPP2, GST–NAPP2(1–203), GSTNAPP2(203–377), and GST, respectively, was probed with 32 P-labeled GST–p45 (Fig. 3C). As shown, p45 bound to GST–NAPP2 and GST–NAPP2(1–203) (Fig. 3C, lanes 5 and 6), but not to GST(203–377) or GST (Fig. 3C, lanes 7 and 8). These results demonstrate that NAPP2 binds to p45 and that the p45-interacting domain in NAPP2 is located within the N-terminal 203-amino-acid region. The domain within p45 interacting with NAPP2 has also been mapped. We tested five different GST fusion polypeptides, GST–p45, GST–p45(1–257), GST–p45(256–373), GST–p45(86–373), and GST–p45D(114–255), for their ability to bind NAPP2 in vitro by far-western blot hybridization (Figs. 4A and 4B). Of these, GST–p45(1–257) contains the acti-

425

Article

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

A

B

FIG. 5. Yeast two-hybrid assay of the interaction between NAPP2(1–203) and p45(1–255) or p45(257–373). Yeast cells were cotransformed with different Gal4 DNA binding domain (pGBT) fusion plasmids and Gal4 activation domain (GAD) fusion plasmids. The cells were streaked onto plates lacking tryptophan and leucine (–Trp, –Leu), or those lacking histidine as well (–Trp, –Leu, –His). Only co-expression of NAPP2(1–203) and p45(1–255) allowed growth on plates lacking histidine. pVA3 and pTD1 were contransformed as a positive control.

FIG. 4. Mapping of p45 domain interacting with NAPP2. (A) Coomassiestained gel showing different GST fusions of p45. (B) Autoradiograph of the gel blot hybridized with 32P-labeled GST–NAPP2. (C) Far-western analysis of the interaction between different GST fusions of p45 and FLAG-tagged NAPP2 expressed in a stably transfected K562 cell line.

far-western assay has provided evidence that a mutation in the PY motif at 79–83 of p45, which disrupts the binding of p45 to WW motifs [8], does not affect the p45–NAPP2 interaction (data not shown). A pull-down assay with different GST–p45 fusions as the baits was carried out with extracts prepared from K562 cells stably transfected with FLAG-tagged, NAPP2 expression plasmid. The extract was first incubated with GST, GST–p45, GST–p45(1–257), GST–p45(256–373), GST–p45(86–373), or GST–p45⌬(114–255). The bound proteins in each incubation reaction were then separated by SDS-PAGE, blotted onto nitrocellulose membrane, and probed with mouse monoclonal antibody against the FLAG epitope. The data (Fig. 4C) correlate well with the conclusions derived above from Figs. 4A and 4B.

vation domain of p45 and GST–p45(256–373) contains the DNA binding and dimerization domain of p45 (Fig. 1A). Nitrocellulose membrane blotted with the fusions and GST was probed with 32P-labeled GST–NAPP2. NAPP2 bound to GST–p45, GST–p45(1–257), and GST–p45⌬(114–255) (Fig. 4B, lanes 2, 3, and 6), but not to GST, GST–p45(256–373), or p45(86–373) (Fig. 4B, lanes 1, 4, and 5). Amino acids 1–114 of p45 contain the sequence interacting with NAPP2, and this NAPP2-interacting sequence may be as localized as within the region of amino acids 1–85 of p45 (Fig. 4). The in vitro result further suggests that eukaryotic protein modification is not required for the p45–NAPP2 interaction. A separate

p45–NAPP2 Interaction in Vivo We used the yeast two-hybrid system to determine whether the p45–NAPP2 interaction occurs in vivo. We first determined that p45 or p45(1–257), but not NAPP2 or NAPP2(1–203), activates transcription as Gal4 fusions in the yeast strain HF7C (data not shown). We then cloned p45(1–257) and p45(256–373) fragments, respectively, into the Gal4-AD vector. pGBT-NAPP2(1-203) was then coexpressed with either pGAD-p45(1-257) or pGAD-p45(256-373) in yeast cells. Gal4-DB–NAPP2(1–203) interacted with Gal4AD–p45(1–257), but not with Gal4-AD or GAL4AD–p45(256–373) (Fig. 5). Together with the p45 pull-down assay of cotransfected K562 cells, we conclude that NAPP2 and p45 interact in vivo.

C

426

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

A

B

NAPP2 as a Repressor of p45-Mediated Activation of Transcription To examine the functional implication of the p45–NAPP2 interaction, we carried out transfection experiments in K562 cells. We first used pSG424 to construct mammalian expression plasmids encoding the yeast Gal4 DNA-binding domain fused either to full-length p45(1–373) or to one of the fragments of p45, that is, 1–83, 1–115, 115–257, or 1–257. In K562 cells, all of the above Gal4-DBD–p45 fusion constructs directed transcriptional activation of a cotransfected CAT reporter cis-linked with five tandem Gal4 binding sites in the plasmid pG5-TK-CAT, albeit with different strengths (data not shown). The effects of coexpression of either NAPP2 or NAPP2(1–203) on the stimulation of CAT expression from pG5-TK-CAT by the various Gal4 fusions of p45 were then tested (Fig. 6A). We cotransfected 250 ng of pSG-p45(1-257) with 10 mg of pMF-NAPP2 or pMF-NAPP2(1-203). Cells were

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

Article

FIG. 6. Transcriptional repression by NAPP2. (A) NAPP2-mediated repression of Gal–p45(1–255)-dependent transcription. NAPP2 or NAPP2(1–203) was tested for the ability to repress the p45 activation domain, that is p45(1–255). The data in lanes 1 and 2 showed the CAT expression from pG5-TK-CAT driven by coexpressed Gal–p45(1–255) in the presence of the pMEP4 vector. Substitution of the pMEP4 vector in the above cotransfection assay with either pMF-NAPP2 or pMF-NAPP2(1-203) inhibited CAT expression, as shown in lanes 3–6. The inhibition was greater when Zn ions were added (compare lanes 4 and 6 with 3 and 5, respectively). No CAT expression was observed when only pG5-TK-CAT and pGBT were used in cotransfection (lane 7). (B) Quantitative comparison of the inhibitory effects of NAPP2 or NAPP2(1–203) on the expression of pG5-TKCAT driven by different Gal–p45 fusions. The assays were carried out similar to (A), and were presented in bar histograms. The reporter expression driven by the activation domains of CTF and VP16 fused to the Gal4 DNA binding domain were included for comparison.

split into two plates soon after transfection, and one set was supplemented with 100 mM ZnSO4. Extracts prepared from cells 48 hours post-transfection were then assayed for the CAT activities. Cotransfection with the pMF vector did not affect CAT expression directed by Gal4-DBD–p45(1–257), in the presence or absence of ZnSO4. However, expression of NAPP2 or NAPP2(1–203) inhibited the Gal4-DBD–p45(1–257) directed CAT expression. The addition of Zn further enhanced the inhibitory effects of both NAPP2 and NAPP2(1–203) (Fig. 6A). It is likely that NAPP2 binds p45(1–257), and thus blocks its contact with the general transcription factors or coactivators assembled at the promoter. Also, although NAPP2(1–203) lacks the putative NLS at amino acids 350–366, the polypeptide is probably small enough to get into the nucleus by diffusion. We then used the same cotransfection approach to test the inhibitory effects of NAPP2 on the abilities of different Gal4DBD–p45 deletions to activate transcription. Coexpression of NAPP2 or NAPP2(1–203) inhibited transcriptional activation by Gal4-DBD fusions of p45(1–83), p45(1–115), p45(1–257), and p45(1–373) (Fig. 6B). However, only moderate inhibition of activation by p45(115–257) was observed. Thus, the domains for the physical interaction between p45 and NAPP2 (Fig. 5) are also required for their functional interaction. Finally, the repressor-like activity of NAPP2 appeared to be p45-specific. Two other activation domains have been tested in the cotransfection assay, the proline-rich activator CTF [24] and the acidic activator VP16 [25]. Neither NAPP2 nor NAPP2(1–203) significantly affected the CAT expression activated by Gal4-DBD fusions of CTF or VP16 (Fig. 6B). These transfection data have provided functional support for a scenario in which NAPP2 has a negative role in p45/NF-E2 activated transcriptional processes. NAPP2 as a Corepressor When Recruited to the Vicinity of Promoter DNA Because NAPP2 or NAPP2(1–203) also inhibited to some extent the transcriptional activation mediated by p45(115–257) and CTF, respectively, which could not bind NAPP2, we wanted to examine whether NAPP2 possesses a p45-independent,

427

Article

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

FIG. 7. Corepressor assay of NAPP2. Gal–NAPP2 functions as a transcriptional repressor when directly bound to the Gal4-binding sites in plasmid pG5-TK-CAT. pSG424 vector (left) and increasing amounts (5 ␮g or 10 ␮g) of pSGNAPP2 and pSG-NAPP2(1-203) were cotransfected, respectively, into K562 cells with 5 ␮g of the reporter plasmid pG5-TK-CAT. Note that CAT expression was greatly increased by coexpression of Gal–VP16 fusion from 2 ␮g of pSGVP16 (lane VP16), indicating the modularity of the reporter expression. The folds of activities shown below are averages of three independent experiments, with the level of the reporter expression driven by the Gal4 DNA binding domain (lane Gal) as 1.

inhibitory domain that could directly repress basal transcription. To test this possibility, the effects of Gal4-DBD–NAPP2 or Gal4-DBD–NAPP2(1–203) on the expression of pG5-TK-CAT reporter plasmid were investigated. Cotransfection of pSGVP16 with pG5-TK-CAT greatly increased the CAT activity, indicating the modularity of the CAT reporter expression (Fig. 7). However, the use of 2 ␮g or 4 ␮g of either pSG-NAPP2 or pSG-NAPP2(1-203) significantly reduced the CAT activity (Fig. 7). This inhibition is not due to a simple occupancy of the Gal4 binding sites on the reporter plasmid by the proteins, as expression of the GAL4-DNA binding domain alone did not repress basal transcription of the CAT reporter (data not shown) [26].

HDAC1. The identification of NAPP2 suggests that p45/NF-E2 could also function as a transcriptional repressor in erythroid and megakaryocytic cells. In general, transcription of eukaryotic genes is regulated by specific sets of DNA-binding transcription factors. The functioning of the DNA-binding factors are mediated through protein–protein interactions involving nonDNA-binding coregulators as well the basal transcription machinery [28,29]. In rare instances, a DNA-binding factor could behave as either an activator or a repressor, depending on the context of the promoter containing the binding motif. Among this class of dually functional DNA-binding factors is YY1. While the activation of transcription involves interactions with TFIIB, TAFs, and the histone acetyltransferase CBP, transcriptional repression by YY1 is mediated by interaction with histone deacetylase HDAC2 [reviewed in 30]. Previously, NF-E2 has been typically characterized and labeled as a transcriptional activator. It seems to be a positive transcriptional regulator for mammalian globin genes and other erythroid megakaryocytic promoters [reviewed

Association of Histone Deacetylase Activity with NAPP2 The data in Fig. 7 suggested that NAPP2 could function as a transcriptional corepressor. To determine whether this function of NAPP2, similar to several other known corepressors, is mediated through histone deacetylases [reviewed in 27], we carried out a pull-down assay of K562 extract using different GST–NAPP2 fusions. Each GST–NAPP2 fusion alone did not have any histone deacetylase activity (Fig. 8). After incubation with the K562 cell lysates, however, GST–NAPP2 or GST–NAPP2(1–203) pulled down proteins exhibiting strong histone deacetylase activities. Neither GST nor GST–NAPP2(203–377) could pull down any histone deacetylase activity. The K562 extract proteins pulled down by NAPP2 were further tested for the presence of the individual histone deacetylases by western blot. GST–NAPP2 is associated with HDAC1, but not with HDAC2 or HDAC3 (Fig. 9).

DISCUSSION By expression cloning, we have identified a protein, NAPP2, that interacts with the erythroid- and megakaryocyte-enriched subunit, p45, of the transcription factor NF-E2. NAPP2 behaved as a negative coregulator of p45/NF-E2 mediated transcriptional activation. Furthermore, NAPP2 appeared to be a novel corepressor capable of recruiting the histone deacetylase

428

FIG. 8. Pull-down assay of HDAC activities. A histone deacetylase assay was carried out on different GST–NAPP2 fusions, with (tubes 1–4) or without (tubes 5–8) prior incubation with K562 whole-cell extract. The beads used contain GST (1 and 5), GST–NAPP2 (2 and 6), GST–NAPP2(1–203) (3 and 7) or GST–NAPP2(203–377) (4 and 8). Sample 9 contained beads only. Values of triplicate samples are presented on the left of the figure. The relative activities are indicated on the right axis.

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

FIG. 9. Complex formation of NAPP2 with HDAC1. GST, GST–NAPP2, or GST–NAPP2(203–377) was incubated with K562 cell lysate as in Fig. 8. The bound proteins in the pull-down fraction (lanes 2–4) as well as the unbound proteins in the post-pull-down supernatants (lanes 5–7) were then resolved on SDSPAGE gel, blotted onto nitrocellulose membranes, and probed with the individual HDAC antibodies. GST–NAPP2 seems to interact only with HDAC1.

Article

Alternatively, NAPP2 could function as a corepressor when recruited to the promoter DNA (Fig. 10, model B). Similar to several other corepressors, including CIR [33], NAPP2 functions as a potent transcription repressor when recruited to promoters through a tethered, heterologous DNA-binding domain such as Gal4 (Fig. 8). NAPP2 joins the group of corepressors that could interact with histone deacetylases or HDACs, either directly like CIR [33] or indirectly in a multiple-protein complex like the nuclear hormone co-receptors NCoR [26] and mSin3 [34]. The targets of deacetylation by HDACs recruited through these corepressors could be the nucleosomes [reviewed in 35], specific transcription activators/coactivators [36,37], or general transcription factors such as TFIIE and TFIIF [reviewed in 29]. Similar to these corepressors, NAPP2 very likely exists in specific complexes containing HDAC1 (Figs. 8 and 9). Although it is unknown whether NAPP2 physically interacts with HDAC1, the latter could be recruited to the vicinity of the promoter through the NAPP2/NF-E2/DNA complex, carry out specific deacetylation reactions, and thus repress transcriptional activation. It is notable that the erythroid-enriched transcription factors GATA-1 and EKLF are both acetylated in vivo [36,37]. Furthermore, the binding sites of these factors are often clustered near the NF-E2 binding motifs within different

in 31]. However, NF-E2 has been implied as a repressor for the expression of a human ␨-globin promoter, through its binding to the HS-40 element [16,32]. In that study, a 1-bp mutation interfering with NF-E2 factor-binding to a NFE2/AP1 motif of HS-40 significantly de-repressed a cislinked human ␨-globin promoter in the fetal and adult erythroid cells of transgenic mice [32]. Although the in vivo involvement of NAPP2 in the ␨-globin promoter silencing is still uncertain at the moment, the identification of this p45interacting, negative transcription coregulator implies that p45/NF-E2 could also function as a DNA-binding repressor in a promoter context- and cellular environment-specific manner. As a negative coregulator, NAPP2 is capable of functioning through one or A both of two different pathways (Fig. 10). In the first pathway, the N-terminal region of NAPP2(1–203) interacts with the activation domain of p45 somewhere between amino acids 1 and 114 (Figs. 3–5). This domain of p45 overlaps with those previously mapped for the interaction B between p45 and other transcriptional factors (Figs. 1A, 6, and 7). Specifically, p45(1–206) binds CBP, a transcription coactivator as well as a histone acetyltransferase [6]. p45(1–80) binds TAF11130/110, a component of the basic transcription machinery [7]. The PY motif at amino acids 89–83 of p45, while not required for NAPP2–p45 interaction, could interact with a subclass of the WW domain-containing proteins, several of which are also transcriptional coactivators [8,17]. Thus, the inhibition of p45-depend- FIG. 10. Models of inhibition of p45-mediated transcription activity by NAPP2. (A) Direct blocking ent transcriptional activation by NAPP2 model. In this model, the transcriptional activation by p45 is inhibited by NAPP2 binding to the acti(Fig. 6) could be accomplished through vation domain of p45. This binding blocks the physical contact between p45 activation domain and other direct blocking of the physical interaction transcriptional activators or coactivators. (B) Corepressor model. In this model, NAPP2 acts as a corebetween p45 and one or more of the above pressor, and it brings in HDAC1. The latter enzyme could remove essential acetyl groups from the nucleosomes near the promoter, a coactivator such as CBP/P300, or a basic transcription factor. coactivators (Fig. 10, model A).

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

429

Article

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

erythroid-specific regulatory elements, including the ␤-globin LCR and the HS-40 enhancer. Thus, erythroid- and/or megakaryocyte-enriched transcription factors such as GATA1 and EKLF might also be the targets of the p45–NAPP2 recruited HDAC(s). NAPP2 has been identified for a while [38]. However, it was also cloned independently as a polypeptide, in human known as NAPP2, required for protein docking onto the peroxisomal membrane [19]. In fact, NAPP2 is conserved in the eukaryotes from yeast to mammals [18–20]. Thus, NAPP2 is a bifunctional protein involved in transcriptional regulation as well as in peroxisomal metabolism. Bifunctional proteins involved in both transcription and membrane localization are not without precedent, for example, the NEDD4 (also known as hRPF1) family of proteins. Human NEDD4 functions as a hormone receptor potentiation factor [39], whereas the mouse homolog has been found to bind epithelial Na+ channels and regulate Na+ channel degradation [40]. Further, sterol regulatory transcription factors SREBP-1 and SREBP-2 are both membrane proteins. Upon sterol simulation, these proteins are proteolytically cleaved and enter into the nucleus to function as transcriptional activators [reviewed in 41]. More recently, another cell-surface protein, amyloid-␤ precursor protein (APP), has been implicated in transcriptional regulation [42]. Whether the two different functional pathways of NAPP2 are also linked to each other remains an intriguing possibility. It should be noted that while most of the NAPP2 molecules are localized on the cytosolic peroxisomes [19], the immunocytological and biochemical methods used could not rule out that minute amounts also reside within the nuclei. Finally, like the SREBPs and APP, NAPP2 protein also undergoes site-specific proteolytic cleavages (Fig. 3C) [20,43], further suggesting mechanistic similarities among the three systems.

MATERIALS AND METHODS Plasmid construction. All recombinant work was essentially carried out as described [44]. Plasmids encoding glutathione-S-transferase (GST) fusions of p45 and of different deletion versions of p45 were constructed in the pGEX-2T vector. The GST–p45 fusion was also constructed in pGEX-2TK. Cloning of GST–p45 has been described [8]. All the plasmids were sequenced for correct reading frames throughout this study. A plasmid encoding GST–p45(1–257) was constructed by ligation of a blunt end, EcoRI-Banl fragment of p45 in blunt Aval site of pGEX-2T. pGST-p45(256-373) was constructed by blunt-end cloning of a BanI-EcoRI, p45 fragment to a blunt SmaI site of pGEX-2T. pGST-p45(86373) and pGST-p45(114-255) were constructed by using the unique restriction sites in p45 cDNA. For pGST-p45(86-373), SacI was used to remove cDNA sequence 5⬘ to codon 86. For pGST-p45⌬(114-255), a StuI-BanI fragment coding for amino acids 114–255 was removed from the p45 cDNA. pBS-NAPP2 was made by blunt-end cloning of a PCR-amplified, 1.9-kb cDNA from the lgt11 clone at the EcoRV site of pBluescript. pGST-NAPP2 was constructed by blunt-end cloning of the EcoRV-PstI fragment of NAPP2 at a blunt EcoRI site of pGEX-2TK. A 600-bp N-terminal fragment was excised from pGSTNAPP2 by BamHI digestion, and cloned at the BamHI site of pGEX-2T. This plasmid was expressed to generate GST–NAPP2(1–203) fusion polypeptide.

430

For yeast two-hybrid screening, the same EcoRV-PstI fragment of NAPP2 from above was ligated at a blunt BamHI site of yeast GAL4-DBD vectors pGBT9 and pAS2-1, respectively. pGBT-NAPP2(1-203) was constructed by excising the NAPP2 C-terminal 1.2-kb sequence from pGBT-NAPP2 with BamHI plus SalI. pGAD-p45(1-257) and pGAD-p45(256-373) were constructed by cloning of appropriate p45 fragments at the EcoRI site of pGAD424. For cotransfection experiments of mammalian cells, we used plasmids pG5-TK-CAT, pSG424, and pSG-VP16 constructed in Mark Ptashne’s lab at Harvard University. FLAG epitope was inserted in the metallothionein promoter vector pMEP4 (Invitrogen) by first ligating an annealed oligo at the KpnI and HindIII sites to create the vector pMF. A PCR-amplified N-terminal fragment of NAPP2 with flanking HindIII and BamHI sites was then cloned at the same sites in pMF to generate pMF-NAPP2(1-203). Finally, pMF-NAPP2 was constructed by cloning of a C-terminal, 1.2-kb BamHI fragment of NAPP2 at the same site in pMF-NAPP2(1-203). Different p45 fragments, NAPP2, and NAPP2(1–203) fragments were also cloned, respectively, in-frame with the Gal4 DNA-binding domain in the mammalian expression vector pSG424 [25]. In particular, an EcoRV-PstI digested NAPP2 fragment cut out from pBS-NAPP2 was blunt-end ligated at the BamHI site of pSG424. To construct pSG-NAPP2(1203), pSG-NAPP2 was digested with (BamHI+SalI) to remove NAPP2(204–377), and a remaining plasmid portion was blunt-end ligated. Expression cloning. GST–p45 fusion protein was expressed in Escherichia coli DH5␣ and purified to homogeneity by GST-agarose as described [8]. The protein was radiolabeled in vitro with bovine heart muscle kinase A catalytic subunit (Sigma) and [␥-32P]ATP. A K562 ␭gt11 cDNA expression library (Clontech) was used for screening as described [45]. Filters were blocked with extracts prepared from GST- expressing E. coli, and then probed with radiolabeled GST–p45 in Hyb75 buffer. Plaques hybridizing with p45 were visualized by autoradiography. Pure plaque expressing NAPP2 was isolated by repeating the screening. Phage lysates were prepared using the E. coli Y1089 infected with NAPP2 phage. SDS-PAGE and far-western blot were done as described [8, 44]. Transient transfection of cell culture. Maintenance of human erythroid K562 cell lines, transient DNA transfection, and CAT assay were as described [46]. DNA-liposome formation was achieved by mixing 3 ␮l lipofectin (Gibco) and 1 ␮g DNA. DNA–liposome complexes were incubated with K562 cells overnight in the absence of antibiotics and serum before replacement with medium containing the serum and antibiotics. At 48 hours after transfection, cells were harvested, washed with 0.25 M Tris-HC1, pH 7.5, and stored at –70⬚C until analysis. All the transfection mixtures contained a control ␤-galactosidase expression plasmid for estimation of the transfection efficiencies. Histone deacetylase assays. For the test of association of NAPP2 with histone deacetylases, K562 extract in PBS buffer with 0.1% NP-40 was prepared as described [47]. GSH beads containing 2 ␮g GST or GST fusions of different proteins in Hyb 75 buffer [45] were incubated with 20 ␮g K562 extract at 4⬚C for 2 hours, and washed as described [34]. The presence of HDACs in the pulldown complexes were examined by standard western blot analysis using different HDAC antibodies. To assay for HDAC activities [48], the complexes from each pull-down reaction were incubated for 60 minutes at 37°C with 500 mg acid-soluble histones isolated from [3H]acetate-labeled chicken erythrocytes. The incubation was terminated by adding acetic acid/HCl to a final concentration of 0.12 N/0.72 N. The released [3H]acetate was extracted with ethyl acetate and quantified by scintillation counting. Each sample was assayed three times, and the non-enzymatic release of label was subtracted to obtain the final values.

ACKNOWLEDGMENTS We thank Ron DePinho (Harvard University) and Mark Ptashne (Sloan Kettering Memorial Hospital) for plasmids. This research was supported by grants from the US National Institutes of Health (NIH DK 29800 to C.-K.J.S., and NIH GM58486 to E.S.), the Canada Research Council (J.R.D.), and the National Science Council, the National Health Research Institute, and the Academia Sinica of Taipei, Taiwan, ROC. RECEIVED FOR PUBLICATION JULY 9; ACCEPTED DECEMBER 27, 2001.

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

doi:10.1006/geno.2002.6714, available online at http://www.idealibrary.com on IDEAL

REFERENCES 1. Andrews, N. C., Erdjument-Bromage, H., Davidson, M. B., Tempst, P., and Orkin, H. S. (1993). Erythroid transcription factor NF-E2 is a haematopoietic-specific basic-leucine zipper protein. Nature 362: 722–728. 2. Blank, V., Kim, M. J., and Andrews, N. C. (1997). Human MafG is a functional partner for p45 NF-E2 in activating globin gene expression. Blood 89: 3925–3935. 3. Igarashi, K., et al. (1994). Regulation of transcription by dimerization of erythroid factor NF-E2 p45 with small Maf proteins. Nature 367: 568–572. 4. Kotkow, K. J., and Orkin, S. H. (1995). Dependence of globin gene expression in mouse erythroleukemia cells on the NF-E2 heterodimer. Mol. Cell. Biol. 15: 4640–4647. 5. Lu, S. J., Rowan, S., Bani, M. R., and Ben-David, Y. (1994). Retroviral integration within the Fli-2 locus results in inactivation of the erythroid transcription factor NF-E2 in Friend erythroleukemias. Proc. Natl. Acad. Sci. USA 91: 8398–8402. 6. Cheng, X., Reginato, M. J., Andrews, N. C., and Lazar, M. A. (1997). The transcriptional integrator CREB-binding protein mediates positive cross talk between nuclear hormone receptors and the hematopoietic bZip protein p45/NF-E2. Mol. Cell. Biol. 17: 1407–1416. 7. Amroli, P. J., et al. (1997). The activation domain of the enhancer binding protein p45NFE2 interacts with TAFII130 and mediates long-range activation of the ␣- and ␤-globin gene loci in an erythroid cell line. Proc. Natl. Acad. Sci. USA 94: 10051–10056. 8. Gavva, N. R., Gavva, R., Ermekova, K., Sudol, M., and Shen, C.-K. J. (1997). Interaction of WW domains with hematopoietic transcription factor p45/NF-E2 and RNA polymerase. J. Biol. Chem. 272: 24105–24108. 9. Armstrong, J. A., and Emerson, B. M. (1996). NF-E2 disrupts chromatin structure at human ␤-globin locus control region hypersensitive site 2 in vitro. Mol. Cell. Biol. 16: 5634–5644. 10. Gong. Q. H., McDowell, J. C., and Dean, A. (1996). Essential role of NF-E2 in remodeling of chromatin structure and transcriptional activation of the ⑀-globin gene in vivo by 5⬘ hypersensitive site 2 of the ␤-globin locus control region. Mol. Cell. Biol. 16: 6055–6064. 11. Martin, D. I., Fiering, S., and Groudine, M. (1996). Regulation of ␤-globin gene expression: straightening out the locus. Curr. Opin. Genet. Dev. 6: 488–495. 12. Jarman, A. P., et al. (1991). Characterization of the major regulatory element upstream of the human ␣-globin gene cluster. Mol. Cell. Biol. 11: 4679–4689. 13. Mignotte, V., Eleouet, J. F., Raich, N., and Romeo, P. H. (1989). Cis- and transacting elements involved in the regulation of the erythroid promoter of the human porphobilinogen deaminase gene. Proc. Natl. Acad. Sci. USA 86: 6548–6552. 14. Deveaux, S., et al. (1997). p45 NF-E2 regulates expression of thromboxane synthase in megakaryocytes. EMBO J. 16: 5654–5661. 15. Daftari, P., Gavva, N. R., and Shen, C.-K. J. (1999). Distinction between AP1 and NF-E2 factor-binding at specific chromatin regions in mammalian cells. Oncogene 18: 5482–5486. 16. Wen, S. -C., et al. (2000). Loading of DNA-binding factors to an erythroid enhancer. Mol. Cell. Biol. 20: 1993–2003. 17. Chen, X., et al. (2001). Human ITCH is a coregulator of the hematopoietic transcription factor NF-E2. Genomics 73: 238–241. 18. Komori, M., et al. (1997). The Hansenula polymorpha PEX14 gene encodes a novel peroxisomal membrane protein essential for peroxisome biogenesis. EMBO J. 16: 44–53. 19. Fransen, M., Terlecky, S. R., and Subramani, S. (1998). Identification of a human PTS1 receptor docking protein directly required for peroxisomal protein import. Proc. Natl. Acad. Sci. USA 95: 8087–8092. 20. Shimizu, N., et al. (1999). The peroxin Pex14p. cDNA cloning by functional complementation on a Chinese hamster ovary cell mutant, characterization, and functional analysis. J. Biol. Chem. 274: 12593–12604. 21. Stros, M., Stokrova, J., and Thomas, J. O. (1994). DNA looping by the HMG-box domains of HMG1 and modulation of DNA binding by the acidic C-terminal domain. Nucleic Acids Res. 22: 1044–1051. 22. Greenberg, S. M., Rosenthal, D. S., Greeley, T. A., Tantravahi, R., and Handin, R. I. (1988). Characterization of a new megakaryocytic cell line: the Dami cell. Blood 72: 1968–1977. 23. Shivdasani, R. A., et al. (1995). Transcription factor NF-E2 is required for platelet for-

GENOMICS Vol. 79, Number 3, March 2002 Copyright © 2002 Elsevier Science (USA). All rights reserved.

24.

25. 26. 27. 28. 29. 30. 31. 32. 33.

34. 35. 36. 37.

38. 39.

40. 41. 42. 43. 44. 45. 46.

47. 48.

Article

mation independent of the actions of thrombopoietin/MGDF in megakaryocyte development. Cell 81: 695–704. Altmann, H., Wendler, W., and Winnacke, E. L. (1994). Transcriptional activation by CTF proteins is mediated by a bipartite low-proline domain. Proc. Natl. Acad. Sci. USA 91: 3901–3905. Sadowski, I., Ma, J., Triezenberg, S., and Ptashne, M. (1988). GAL4-VP16 is an unusually potent transcriptional activator. Nature 335: 563–564. Alland, L., et al. (1997). Role for N-CoR and histone deacetylase in Sin3-mediated transcriptional repression. Nature 387: 49–55. Kuo, M. H., and Allis, C. D. (1998). Roles of histone acetyltransferases and deacetylases in gene regulation. Bioessays 20: 615–626. Mannervik, M., Nibu, Y., Zhang, H., and Levine, M. (1999). Transcriptional coregulators in development. Science 284: 606–609. Imhof, A., and Wolffe, A. P. (1998). Transcription: gene control by targeted histone acetylation. Curr. Biol. 8: R422–424. Thomas M. J., and Seto, E. (1999). Unlocking the mechanisms of transcription factor: are chromatin modifying enzymes the key? Gene 236: 197–208. Orkin, S. H. (1995). Regulation of globin gene expression in erythroid cells. Eur. J. Biochem. 231: 271–281. Huang, B. -L., et al. (1998). Derepression of human embryonic ␨ globin promoter by a locus control region sequence. Proc. Natl. Acad. Sci. USA 95: 14669–14674. Hsieh, J. J., Zhou, S., Chen, L., Young, D. B., and Hayward, S. D. (1999). CIR, a corepressor linking the DNA binding factor CBF1 to the histone deacetylase complex. Proc. Natl. Acad. Sci. USA 96: 23–28. Laherty, C. D., et al. (1997). Histone deacetylases associated with the mSin3 corepressor mediate mad transcriptional repression. Cell 89: 349–356. Kuo, M. H., and Allis, C. D. (1998). Role of histone acetyltransferase and deacetylase in gene regulation. Bioassays 20: 615–626. Boyes, J., Byfield, P., Nakatani, Y., and Ogryzko, V. (1998). Regulation of activity of the transcription factor GATA-1 by acetylation. Nature 396: 594–598. Zhang, W., and Bieker, J. J. (1998). Acetylation and modulation of erythroid Kruppellike factor (EKLF) activity by interaction with histone acetyltransferases. Proc. Natl. Acad. Sci. USA 95: 9855–9860. Gavva, N. R., Daftari, P., L. Yang, L. P., and Shen, C.-K. J. (1998). Mechanism of transcription 63rd annual meeting, p. 54. Cold Spring Harbor Laboratory, New York. Imhof, M. O., and McDonnell, D. P. (1996). Yeast RSP5 and its human homolog hRPF1 potentiate hormone-dependent activation of transcription by human progesterone and glucocorticoid receptors. Mol. Cell. Biol. 16: 2594–2605. Dinudom, A., et al. (1998). Nedd4 mediates control of an epithelial Na+ channel in salivary duct cells by cytosolic Na+. Proc. Natl. Acad. Sci. USA 95: 7169–7173. Brown, M. S., and Goldstein, J. L. (1997). The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 89: 331–340. Cao, X., and Sudhof, T. C. (2001). A transcriptively active complex of APP with Fe65 and histone acetyltransferase Tip60. Science 293: 115–120. Albertini, M., et al. (1997). Pex14p, a peroxisomal membrane protein binding both receptors of the two PTS-dependent import pathways. Cell 89: 83–92. Sambrook, J., Fritsch, E. F. M., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Plainview, NY. Kaelin, W. G., et al. (1992). Expression cloning of a cDNA encoding a retinoblastomabinding protein with E2F-like properties. Cell 70: 351–364. Zhang, Q., Rombel, I., Reddy, G. N., Gang, J. B., and Shen, C.-K. J. (1995). Functional roles of in vivo footprinted DNA motifs within an ␣-globin enhancer. Erythroid lineage and developmental stage specificities. J. Biol. Chem. 270: 8501–8505. Blackwood, E. M., and Eisenman, R. N. (1991). Max: a helix-loop-helix zipper protein that forms a sequence-specific DNA binding complex with Myc. Science 251: 1211–1217. Hendzel, M. J., Delcuve, G. P., and Davie, J. R. (1991). Histone deacetylase is a component of the internal nuclear matrix. J. Biol. Chem. 266: 21936–21942.

431