DNA Binding and Gene Activation Properties of the Nmp4 Nuclear Matrix Transcription Factors
2002; Elsevier BV; Volume: 277; Issue: 18 Linguagem: Inglês
10.1074/jbc.m107496200
ISSN1083-351X
AutoresKitti Torrungruang, Marta Alvarez, Rita Shah, Jude E. Onyia, Simon J. Rhodes, Joseph P. Bidwell,
Tópico(s)Genomics and Chromatin Dynamics
ResumoSplice variants of the Nmp4gene include nuclear matrix transcription factors that regulate the type I collagen α1(I) polypeptide chain (COL1A1) promoter and several matrix metalloproteinase (MMP) genes. To date, these are the only Cys2His2 zinc finger proteins known to bind within the minor groove of homopolymeric (dA·dT) DNA. Nmp4 isoforms contain from 5 to 8 Cys2His2 zinc fingers, an SH3-binding domain that overlaps with a putative AT-hook and a polyglutamine-alanine repeat (poly(QA)). To determine the mechanistic significance of Cys2His2 zinc finger association with this unusual consensus DNA binding element, we identified the Nmp4 DNA-binding and transcriptional activation domains. Zinc fingers 2, 3, and 6 mediated association with the homopolymeric (dA·dT) COL1A1/MMP DNA consensus element. The N terminus of the Nmp4 protein exhibited a strongtrans-activation capacity when fused to the GAL4 DNA-binding domain, but this activity was masked within the context of the full-length Nmp4-GAL4 DNA-binding domain chimera. However, upon binding to the COL1A1/MMP homopolymeric (dA·dT) element, the native Nmp4 protein up-regulated transcription, and the poly(QA) domain acquired a significant role intrans-activation. We propose that allosteric effects induced upon zinc finger association with the homopolymeric (dA·dT) minor groove confer context-specific functionality to this unusual family of Cys2His2 transcription factors. Splice variants of the Nmp4gene include nuclear matrix transcription factors that regulate the type I collagen α1(I) polypeptide chain (COL1A1) promoter and several matrix metalloproteinase (MMP) genes. To date, these are the only Cys2His2 zinc finger proteins known to bind within the minor groove of homopolymeric (dA·dT) DNA. Nmp4 isoforms contain from 5 to 8 Cys2His2 zinc fingers, an SH3-binding domain that overlaps with a putative AT-hook and a polyglutamine-alanine repeat (poly(QA)). To determine the mechanistic significance of Cys2His2 zinc finger association with this unusual consensus DNA binding element, we identified the Nmp4 DNA-binding and transcriptional activation domains. Zinc fingers 2, 3, and 6 mediated association with the homopolymeric (dA·dT) COL1A1/MMP DNA consensus element. The N terminus of the Nmp4 protein exhibited a strongtrans-activation capacity when fused to the GAL4 DNA-binding domain, but this activity was masked within the context of the full-length Nmp4-GAL4 DNA-binding domain chimera. However, upon binding to the COL1A1/MMP homopolymeric (dA·dT) element, the native Nmp4 protein up-regulated transcription, and the poly(QA) domain acquired a significant role intrans-activation. We propose that allosteric effects induced upon zinc finger association with the homopolymeric (dA·dT) minor groove confer context-specific functionality to this unusual family of Cys2His2 transcription factors. type I collagen α1(I) polypeptide chain metalloproteinase electrophoretic mobility shift analysis Src homology 3 dimethyl sulfate DNA-binding domain activation domain amino acid nucleotide analysis of variance The Cys2His2 zinc finger proteins comprise the largest group of eukaryotic DNA-binding proteins (1Iuchi S. Cell. Mol. Life Sci. 2001; 58: 625-635Crossref PubMed Scopus (301) Google Scholar, 2Wolfe S.A. Nekludova L. Pabo C.O. Annu. Rev. Biophys. Biomol. Struct. 2000; 29: 183-212Crossref PubMed Scopus (787) Google Scholar). This type of zinc finger contains two cysteine and two histidine residues that bind a single zinc atom stabilizing the finger, a loop consisting of a conserved ββα structure (2Wolfe S.A. Nekludova L. Pabo C.O. Annu. Rev. Biophys. Biomol. Struct. 2000; 29: 183-212Crossref PubMed Scopus (787) Google Scholar, 3Dreier B. Beerli R.R. Segal D.J. Flippin J.D. Barbas III, C.F. J. Biol. Chem. 2001; 276: 29466-29478Abstract Full Text Full Text PDF PubMed Scopus (264) Google Scholar). The amino acids on the surface of the α-helix contact bases in the DNA major groove (2Wolfe S.A. Nekludova L. Pabo C.O. Annu. Rev. Biophys. Biomol. Struct. 2000; 29: 183-212Crossref PubMed Scopus (787) Google Scholar). This motif is repeated in tandem and allows for binding to DNA sequences of variable lengths (3Dreier B. Beerli R.R. Segal D.J. Flippin J.D. Barbas III, C.F. J. Biol. Chem. 2001; 276: 29466-29478Abstract Full Text Full Text PDF PubMed Scopus (264) Google Scholar, 4Segal D.J. Dreier B. Beerli R.R. Barbas III, C.F. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 2758-2763Crossref PubMed Scopus (377) Google Scholar). The modular design of the zinc finger domain also affords many combinatorial possibilities for specific DNA recognition underlying the tremendous variety of consensus sequences recognized by zinc finger proteins (3Dreier B. Beerli R.R. Segal D.J. Flippin J.D. Barbas III, C.F. J. Biol. Chem. 2001; 276: 29466-29478Abstract Full Text Full Text PDF PubMed Scopus (264) Google Scholar, 4Segal D.J. Dreier B. Beerli R.R. Barbas III, C.F. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 2758-2763Crossref PubMed Scopus (377) Google Scholar, 5Zhang L. Spratt S.K. Liu Q. Qi B. Johnstone H. Raschke E.E. Jamieson A. Rebar E.J. Wolffe A.P. Case C.C. J. Biol. Chem. 2000; 275: 33850-33860Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar). Despite the vast DNA recognition potential of Cys2His2 zinc finger proteins, few are known to associate with AT-rich DNA. These DNA sequences can act as significant structural elements in transcription (6Reeves R. Environ. Health Perspect. 2000; 108: 803-809Crossref PubMed Google Scholar, 7Koch K.A. Thiele D.J. J. Biol. Chem. 1999; 274: 23752-23760Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar, 8Nelson H.C. Finch J.T. Luisi B.F. Klug A. Nature. 1987; 330: 221-226Crossref PubMed Scopus (921) Google Scholar). The narrow minor groove of AT-rich DNA furnishes a structural recognition motif for many architectural transcription factors, proteins that bend DNA and influence transcription by altering the interactions between othertrans-acting proteins (6Reeves R. Environ. Health Perspect. 2000; 108: 803-809Crossref PubMed Google Scholar). Matrix attachment regions are typically long stretches (200–800 bp) of AT-rich DNA that anchor chromatin loops to the nuclear matrix and provide chromatin boundaries and binding sites for nuclear matrix regulatory proteins (9Bode J. Benham C. Knopp A. Mielke C. Crit. Rev. Eukaryotic Gene Expression. 2000; 10: 73-90Crossref PubMed Google Scholar). Homopolymeric (dA·dT) DNA can have functional significance apart from interaction with trans-acting proteins (7Koch K.A. Thiele D.J. J. Biol. Chem. 1999; 274: 23752-23760Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar, 8Nelson H.C. Finch J.T. Luisi B.F. Klug A. Nature. 1987; 330: 221-226Crossref PubMed Scopus (921) Google Scholar). By comparison to B-form DNA, the homopolymeric (dA·dT) sequences have a shorter helical repeat (10.0 bp/turn as opposed to 10.5 bp/turn), a narrower minor groove (∼9 Å as compared with ∼15 Å), and additional bifurcated hydrogen bonds that provide added structural stability (7Koch K.A. Thiele D.J. J. Biol. Chem. 1999; 274: 23752-23760Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar, 8Nelson H.C. Finch J.T. Luisi B.F. Klug A. Nature. 1987; 330: 221-226Crossref PubMed Scopus (921) Google Scholar). The rigidity of these elements resists conforming tightly to the face of the nucleosome, creating localized DNA distortions on either end of the element that provides “access windows” for trans-acting factors (7Koch K.A. Thiele D.J. J. Biol. Chem. 1999; 274: 23752-23760Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar, 10Lascaris R.F. Groot E. Hoen P.B. Mager W.H. Planta R.J. Nucleic Acids Res. 2000; 28: 1390-1396Crossref PubMed Google Scholar, 11Suter B. Schnappaf G. Thoma F. Nucleic Acids Res. 2000; 28: 4083-4089Crossref PubMed Google Scholar, 12Iyer V. Struhl K. EMBO J. 1995; 14: 2570-2579Crossref PubMed Scopus (348) Google Scholar). The Nmp4 proteins (Fig. 1A) are the only known Cys2His2 zinc finger proteins that bind to the minor groove of homopolymeric (dA·dT) sites, such as those present in the promoters of many extracellular matrix genes including the type I collagen α1(I) polypeptide chain (COL1A1)1 promoter and several matrix metalloproteinases (MMP) (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar, 14Nakamoto T. Yamagata T. Sakai R. Ogawa S. Honda H. Ueno H. Hirano N. Yazaki Y. Hirai H. Mol. Cell. Biol. 2000; 20: 1649-1658Crossref PubMed Scopus (104) Google Scholar, 15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar) (Fig.1B). The MMP Nmp4 consensus element is embedded within both COL1A1 sites A and B in the reverse orientation (Fig. 1B). Some Nmp4 binding sites in the MMP7promoter contain a stretch of 7–9 As similar to the stretch of 9 Ts in the COL1A1 consensus sequence. In a previous study, we compared the capacities of the Nmp4 isoforms 11H, 13H, 21H, and 28H (GenBankTM accession numbersAF216804, AF216805, AF216806, AF216807, and AB019281, see Fig.1A) to bind to the COL1A1/MMP regulatory sequence (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar). The Nmp4 isoforms 11H, 13H, and 21H are full-length cDNAs encoding in-frame splice variants containing from 6–8 Cys2His2 zinc fingers. Nmp4/11H and Nmp4/13H have eight zinc fingers, whereas Nmp4/21H has six fingers, missing numbers 4 and 5. Nmp4/28H has only the first five zinc fingers, is lacking the poly(QA) domain and C terminus common to the other isoforms, and has an insert just before the first finger (Fig.1A). Using electrophoretic mobility shift analysis (EMSA), the Nmp4 proteins 11H, 13H, and 21H exhibited binding to theCOL1A1/MMP consensus site and yielded binding profiles similar to that obtained with the nuclear matrix fraction from osteoblast-like cells (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar). However, the Nmp4/28H clone did not bind to the consensus element (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar). Therefore, Nmp4/21H is a native truncated isoform exhibiting COL1A1/MMP DNA binding. Nmp4 proteins have an SH3-binding domain (14Nakamoto T. Yamagata T. Sakai R. Ogawa S. Honda H. Ueno H. Hirano N. Yazaki Y. Hirai H. Mol. Cell. Biol. 2000; 20: 1649-1658Crossref PubMed Scopus (104) Google Scholar) that overlaps with a putative AT-hook motif (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar), another potential DNA-binding domain (Fig.1A). The AT-hook is characteristic of the High Mobility Group A superfamily of architectural transcription factors and can mediate binding to the minor groove of AT-rich DNA (6Reeves R. Environ. Health Perspect. 2000; 108: 803-809Crossref PubMed Google Scholar, 16Bustin M. Trends Biochem. Sci. 2001; 26: 152-153Abstract Full Text Full Text PDF PubMed Google Scholar, 17Aravind L. Landsman D. Nucleic Acids Res. 1998; 26: 4413-4421Crossref PubMed Scopus (383) Google Scholar). However, both the SH3-binding and AT-hook domains can mediate protein-protein interactions in the formation of higher order protein complexes in the cytoplasm and nucleus, respectively (18Chin M.T. Pellacani A. Wang H. Lin S.S. Jain M.K. Perrella M.A. Lee M.E. J. Biol. Chem. 1998; 273: 9755-9760Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar, 19Currie R.A. J. Biol. Chem. 1997; 272: 30880-30888Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar, 20Musacchio A. Wilmanns M. Saraste M. Prog. Biophys. Mol. Biol. 1994; 61: 283-297Crossref PubMed Scopus (144) Google Scholar). In the N-terminal region, Nmp4 proteins also possess a serine-threonine-rich motif (amino acids 51–173), similar to trans-activation domains in immunoglobulin transcription factor-1, immunoglobulin transcription factor-2, and Pax6 (21Seipel K. Georgiev O. Schaffner W. EMBO J. 1992; 11: 4961-4968Crossref PubMed Scopus (301) Google Scholar, 22Tang H.K. Singh S. Saunders G.F. J. Biol. Chem. 1998; 273: 7210-7221Abstract Full Text Full Text PDF PubMed Scopus (79) Google Scholar). Additionally, the Nmp4 isoforms include a poly(Q) motif encoded by a CAG trinucleotide repeat and a poly(QA) repeat (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar, 14Nakamoto T. Yamagata T. Sakai R. Ogawa S. Honda H. Ueno H. Hirano N. Yazaki Y. Hirai H. Mol. Cell. Biol. 2000; 20: 1649-1658Crossref PubMed Scopus (104) Google Scholar). Both of these motifs have been shown to influence the transcriptional activity of other keytrans-acting proteins including Cbfa1/Osf2, CA150, and the human androgen receptors (23Beilin J. Ball E.M. Favaloro J.M. Zajac J.D. J. Mol. Endocrinol. 2000; 25: 85-96Crossref PubMed Scopus (216) Google Scholar, 24Suñé C. Hayashi T. Liu Y. Lane W.S. Young R.A. Garcia-Blanco M.A. Mol. Cell. Biol. 1997; 17: 6029-6039Crossref PubMed Scopus (83) Google Scholar, 25Thirunavukkarasu K. Mahajan M. McLarren K.W. Stifani S. Karsenty G. Mol. Cell. Biol. 1998; 18: 4197-4208Crossref PubMed Google Scholar). Here, our objective was to identify the DNA-binding and trans-activation domains of the Nmp4 proteins. We observed that only three specific zinc fingers (2, 3, and 6 of isoform 21H) were required for binding to the homopolymeric (dA·dT) COL1A1/MMP Nmp4 consensus element.Trans-activation experiments revealed a sensitivity of specific transcriptional regulatory domains to their attached DBD or to their DNA-binding state consistent with the apparent multiple functions of the proteins themselves. Human embryonic kidney 293T cells (ATCC, Manassas, VA) were maintained in Dulbecco’s modified Eagle’s medium (Invitrogen) low glucose medium supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin, 25 μg/ml amphotericin, 2 mml-glutamine (Invitrogen), and 10% fetal bovine serum (Sigma). The rat osteosarcoma cells UMR 106-01, a generous gift from Dr. Nicola Partridge (St. Louis University, St. Louis, MO) were grown as described previously (15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar). All cells were maintained in humidified 95% air/5% CO2 at 37 °C. The cDNAs of the wild-type Nmp4 isoforms 11H, 13H, 21H, and 28H were subcloned into pcDNA-3TM (Invitrogen) as described previously (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar). The deletions from N- or C-terminal were generated by cloning the PCR products of truncated 11H or 21H clones into pcDNA-3TM vectors. The internal deletions were prepared using ExSiteTM PCR-based site-directed mutagenesis kit (Stratagene, La Jolla, CA). The integrity of all plasmids was confirmed by DNA sequencing and restriction enzyme digestion. The Nmp4 pcDNA-3 constructs were expressed in vitro using the TNT® Quick-coupled transcription/translation system following the manufacturers’ instructions (Promega, Madison, WI). The molecular weights of the protein products were confirmed by Western analysis. Protein-DNA interactions were characterized using EMSA as described previously (15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar). The 20-μl binding reactions included 75 mm KCl, 15% glycerol, 0.15 mm EDTA, 500 ng of poly(dI·dC), 0.1 mm dithiothreitol, 19 mmHepes (pH 7.5), 0.0075% Nonidet P-40, 2–4 μl of TNT lysate protein or 2 μg of nuclear matrix protein, and 0.5 nm end-labeled rat COL1A1 5′-regulatory fragment (site A = −3489/−3434 nt and site B = −1594/−1541 (15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar)) as the probe. These fragments also serve as Nmp4-MMP binding sites, because the only difference is the orientation of the homopolymeric sequence within the promoters (13, 14, also see Fig. 1B). The rat COL1A1 5′-regulatory region (−3518/+115 nt) was a generous gift from A. Lichtler, B. Kream, and D. Rowe (The University of Connecticut Health Center, Framington, CT). We used Western blot analysis to relatively quantify the amount of each protein in thein vitro transcription translation lysate. The volume of the lysate used in the EMSA (2–4 μl) was based on the data from the Western analysis. Distamycin A hydrochloride and methyl green (Sigma) were prepared as aqueous stock solutions (10 mm). These compounds were added to some of the EMSA-binding reactions and incubated for 25–30 min at room temperature prior to electrophoresis, which was performed on 8% polyacrylamide gels (80:1 acrylamide,N,N′-methylbisacrylamide) in 1 × Tris/glycine EDTA buffer (4 °C). For DMS interference fingerprinting analysis (26Shaw P.E. Stewart A.F. Methods Mol. Biol. 2001; 148: 221-227PubMed Google Scholar, 27He X. Hohn T. Futterer J. J. Biol. Chem. 2000; 275: 11799-11808Abstract Full Text Full Text PDF PubMed Scopus (22) Google Scholar), the ratCOL1A1 5′-regulatory region containing site A (−3518/−3406 nt) was singly end-labeled at either the 5′ or 3′ end. This DNA was then partially methylated by adding 2 μl of dimethyl sulfate (Sigma) for 1 min in 200 μl of reaction buffer (60 mm NaCl, 10 mm Tris-HCl (pH 8.0), 10 mm MgCl2, and 1 mm EDTA). Methylation was stopped with the addition of 0.3 m sodium acetate (pH 7.0), and 200 mm2-mercaptoethanol. The DNA was then precipitated with ethanol, washed, lyophilized, and resuspended in Tris-EDTA buffer. The partially methylated DNA was used as a probe in EMSA with the in vitrotranslated Nmp4/11H and Nmp4/21H proteins as described above. The protein-DNA complexes and unbound probe were excised and eluted from the gel. The bound proteins were removed by extraction with phenol-chloroform, and the DNA was purified using the Qiagen PCR purification kit (Qiagen, Inc, Valencia, CA) followed by ethanol precipitation. To cleave the modified A and G residues, the DNA pellet was resuspended in 10 mm sodium phosphate (pH 6.8)/1 mm EDTA and incubated for 15 min at 92 °C followed by the addition of 100 mm NaOH and incubated for another 30 min at 92 °C. The DNA was then ethanol-precipitated and resuspended in Tris-EDTA buffer; equal counts were lyophilized, dissolved in formamide gel-loading dye, and resolved in a 8% sequencing gel. The EMSA and DMS interference gels were dried under vacuum at 80 °C for 1 h and then at room temperature for 30 min. Kodak XAR film was exposed to the gel with an intensifying screen overnight at −80 °C. GAL4 fusion constructs of full-length and truncated derivatives of Nmp4/21H were obtained by ligating the PCR-generated coding sequences downstream and in frame with a sequence coding for the GAL4-DNA-binding domain (GAL4-DBD, amino acids 1–147) in the pBind vector (Checkmate Mammalian two-hybrid system, Promega). The integrity of all constructs was verified by DNA sequencing and restriction enzyme digestion, and their expression was confirmed by Western analysis. For analysis of activation domains, pFRluc vector (Stratagene) containing five copies of the GAL4 binding element and a basic promoter element (TATA box) upstream of a luciferase gene was used as a reporter. GAL4 fusion constructs and pFRluc reporter were used in 1:2 microgram ratio for transient transfection in 293T cells. The human kidney 293T cells were seeded into 6-well plates at 1.0–1.5 × 105 cells/well. After 24 h, the cells in each well were transiently transfected with DNA using the CalPhos system (CLONTECH, Palo Alto, CA). The transfected cells were harvested with the reporter lysis buffer (Promega) 48 h after transfection. Cells were washed two times with phosphate-buffered saline (137 mm NaCl, 2.7 mm KCl, 4.3 mm Na2HPO47H2O, and 1.4 mm KH2PO4 (pH 7.1)) and then exposed to lysis buffer for 15 min at room temperature. The lysates were processed for luciferase activity according to the manufacturers’ instructions (Promega) by adding 100 μl of Bright-GloTMreagent to 100 μl of cell lysate and reading immediately on a Packard bioscience fusion luminometer (Packard Instrument Company, Meriden, CT). Total protein concentrations were determined using the Coomassie Blue Plus protein assay reagent (Pierce). Results for the luciferase assay were reported as relative light units per microgram of total protein. The human MMP7 promoter (−2300/+30 nt) kindly provided by Dr. Lynn Matrisian (Vanderbilt University, Nashville, TN) was subcloned into the pGL-3 Basic (Promega) using BamHI andXhoI. The sequence fidelity of the MMP7/pGL-3 construct was verified by restriction digests. To assay MMP7activation, the Nmp4/21H or its deletion constructs in the pcDNA-3 expression vector and the MMP7 reporter construct were used in a 1:2 microgram ratio for transient transfection in 293T cells. The luciferase assay was then performed as described above. Nuclear matrix proteins were extracted from UMR 106-01 cells using a sequential extraction protocol as described previously (28Alvarez M. Onyia J. Hock J. Xu H. Long W. Bidwell J. Endocrinology. 1997; 138: 482-489Crossref PubMed Scopus (47) Google Scholar). A one-way analysis of variance (ANOVA) was used to determine the differences among groups with a predeterminedp value of <0.05 to achieve statistical significance. Tukey’s test for post hoc comparison was applied based on the ANOVA results. The truncated Nmp4/28H isoform has an insert before the first zinc finger (Fig.1A). Removal of this insert failed to confer DNA binding to this isoform (data not shown). Therefore, we used Nmp4/21H to characterize the minimum Nmp4-COL1A1/MMP DNA-binding domain, because it is the shortest naturally occurring isoform that binds to these regulatory sites (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar, 14Nakamoto T. Yamagata T. Sakai R. Ogawa S. Honda H. Ueno H. Hirano N. Yazaki Y. Hirai H. Mol. Cell. Biol. 2000; 20: 1649-1658Crossref PubMed Scopus (104) Google Scholar). Wild-type Nmp4/21H and derivatives were subcloned into pcDNA-3, expressed in vitro, and characterized by EMSA using the COL1A1 site A as the probe (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar). The full-length Nmp4/21H isoform is comprised of 518 amino acids (aa), which include zinc fingers 1–3 and 6–8 (Fig.2A, construct 1). The deletion of the C-terminal region (construct 2), the poly(QA) domain (construct 3), or the AT-hook/SH3-binding domain (construct 7) did not abrogate DNA binding to theCOL1A1/MMP consensus sequence (Fig. 2,A and B). The removal of larger regions of the Nmp4/21H protein including the N terminus (construct 5) and the C terminus plus the poly(QA) repeat (construct 4) had no impact on DNA binding as determined by EMSA (Fig. 2, A andB). Deletion of the Cys2His2 zinc fingers from Nmp4/21H (construct 8) abrogated DNA binding (Fig. 2,A and B). Conversely, this domain alone (construct 9) exhibited binding to the consensus sequence (Fig. 2, A and B). To determine whether any specific zinc fingers conferred DNA binding to theCOL1A1/MMP binding site, we deleted different combinations of three, two, or single zinc fingers. We found that zinc fingers 1, 7, and 8 are not required, whereas the combination of zinc fingers 2, 3, and 6 were necessary for mediating Nmp4/21H binding to theCOL1A1/MMP consensus elements. Identical results were obtained using COL1A1 site B (data not shown). Nmp4/28H is missing zinc finger 6, which would explain its inability to bind to theCOL1A1/MMP consensus element (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar). To determine whether the Cys2His2 zinc fingers alone mediated binding to the minor groove of the AT-rich Nmp4 consensus sequence, we used EMSA in combination with distamycin and methyl green. Distamycin is used as a competitor for protein binding to the minor groove of AT-rich DNA (29Abu-Daya A. Fox K.R. Nucleic Acids Res. 1997; 25: 4962-4969Crossref PubMed Scopus (42) Google Scholar), and methyl green competes for the occupation of the major groove of DNA (30Kim S.K. Norden B. FEBS Lett. 1993; 315: 61-64Crossref PubMed Scopus (137) Google Scholar). The binding of the full-length construct of Nmp4/21H to site A of the COL1A1promoter was sensitive to the presence of distamycin (Fig.3A) as demonstrated previously (13Thunyakitpisal P. Alvarez M. Tokunaga K. Onyia J.E. Hock J. Ohashi N. Feister H. Rhodes S.J. Bidwell J.P. J. Bone Miner. Res. 2001; 16: 10-23Crossref PubMed Scopus (56) Google Scholar, 15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar). The deletion of the AT-hook/SH3-binding domain (construct 7) did not alter the effect of distamycin on Nmp4/21H-COL1A1 binding (Fig. 3A). The zinc finger domain alone (construct 9) exhibited the same sensitivity to distamycin as the full-length protein (Fig.3A). Interestingly, the full-length Nmp4/21H construct exhibited a weak sensitivity to the same concentrations of methyl green (Fig. 3B), yet this sensitivity was attenuated upon removal of the putative AT-hook/SH3-binding domain. Finally, DNA binding by the zinc finger domain alone showed no sensitivity to the presence of methyl green (Fig. 3B). In a previous study, we used methidium-propyl EDTA footprinting to demarcate the perimeter of the nuclear matrix extract binding activity along site A of the COL1A1 5′-regulatory region (15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar). Here, we used DMS interference fingerprinting to further delimit the Nmp4 protein-DNA contacts and to compare Nmp4/11H and Nmp4/21H association with this consensus element (Fig. 4). The Nmp4/11H isoform contains eight zinc fingers, whereas the Nmp4/21H has six. Both the Nmp4/11H and Nmp4/21H fingerprints extended across nucleotides −3463/−3458 (Fig. 4), consistent with the methidium-propyl EDTA footprint of the nuclear matrix extract obtained in our earlier study (15Alvarez M. Thunyakitpisal P. Morrison P. Onyia J. Hock J. Bidwell J.P. J. Cell. Biochem. 1998; 69: 336-352Crossref PubMed Scopus (29) Google Scholar). The sequence of the opposing strand in the area of the Nmp4 binding site contains only Cs and Ts and therefore was not suitable for methylation interference analysis (data not shown). We used Nmp4-GAL4-DBD chimeras consisting of either the full-length Nmp4/21H protein or its truncated derivatives fused to the GAL4-DBD to identify domains exhibiting autonomous trans-activation capacity. This was accomplished by co-transfecting 293T cells with the chimeras and a reporter construct driven by a basic promoter linked to five copies of the GAL 4 consensus sequence (see “Materials and Methods”). The N terminus (1–187 aa) of Nmp4/21H exhibited strongtrans-activation capacity and increased promoter activity 20–30-fold over base-line promoter activity (Fig.5). The poly(QA) (396–453 aa) motif lacked trans-activation activity, but in combination with the C terminus, (396–518 aa) showed a weak trans-activation capacity increasing promoter activity ∼2-fold over base line (Fig.5). Interestingly, the strong trans-activation activity of the N terminus was completely masked within the context of the full-length Nmp4/21H-GAL4-DBD chimera (Fig. 5). In fact, the full-length Nmp4-GAL4-DBD chimera (1–518 aa) and the zinc finger-GAL4-DBD chimera (231–395 aa) repressed promoter activity 5- and 9-fold below base line, respectively (Fig. 5). The portion of the Nmp4 protein (188–230 aa) that includes the AT-hook/SH3-binding domain significantly attenuated the strong trans-activation capacity of the N terminus (1–187 aa) but did not completely inhibit it (2.1-fold over base line), suggesting that the zinc finger domain contributed to this masking effect (Fig. 5). Promoter-reporter constructs lacking the GAL4 DNA binding sites did not respond to the various Nmp4/21H-GAL4-DBD chimeric constructs (data not shown), indicating the absence of background effects caused by the zinc finger-containing constructs targeting other sites. To determine whether the Nmp4/21H regulatory domains behave as autonomous modules or instead are sensitive to the identity of their attached DBD or to their DNA binding state, we removed the GAL4-DBD. Kidney 293T cells were co-transfected with a human MMP7 promoter-reporter construct containing the homopolymeric (dA·dT) Nmp4 consensus sequence and an expression vector containing either the full-length Nmp4/21H isoform or one of its truncated derivatives. Full-length Nmp4/21H isoform (1–518 aa) up-regulated the basal activity of the MMP7/luciferase promoter-report
Referência(s)