Genetic Selection for Modulators of a Retinoic-Acid-Responsive Reporter in Human Cells
Deltagen Proteomics, Salt Lake City, Utah 84108(7{-2e., 百拇医药
ABSTRACT(7{-2e., 百拇医药
We used a genetic screening methodology, a human cell line bearing a retinoic-acid-responsive enhanced GFP reporter, and a flow sorter to recover dominant modulators of reporter expression. Four inducers and three suppressors that were fused to the C terminus of a protein scaffold for stability were isolated and their mechanisms of action studied. Mutagenesis experiments indicated that six of these dominant agents exerted their effects at the protein level. The single cDNA coding fragment that was isolated comprised the central 64-amino-acid section of human cyclophilin B, which contained its peptidyl-prolyl isomerase domain; this cyclophilin fragment repressed expression of the retinoic-acid-responsive reporter. The remaining clones encoded peptides shorter than 30 amino acids unrelated to known gene open reading frames. Genetic epistasis studies between the strongest inducer, R3, and a dominant-negative mutant of RAR{alpha} suggest that the two factors function in the same pathway. Transcript microarray analyses suggest that R3 induced a subset of the retinoid-responsive genes in melanoma cells. Finally, yeast two-hybrid assays and co-immunoprecipitation studies of human cell extracts identified PAT1 as a protein that interacts with R3.
DOMINANT genetics has developed over the past several years in an effort to deal with some of the difficulties of screens and selections in cultured mammalian cells, including diploidy and population heterogeneity (HERSKOWITZ 1987 ; KAMB and TENG 2000 ). Using expressed RNA, protein fragments, or peptides (collectively called dominant agents or modulators henceforth), dominant phenotypic modulators have been isolated in several screens, nearly all of which involve selection of survivors from a population of dying or growth-arrested cells (HOLZMAYER et al. 1992 ; GUDKOV et al. 1994 ; AXENOVICH et al. 1998 ; CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; LEVENSON et al. 1999 ; NORMAN et al. 1999 ; XU et al. 1999 ; WELCH et al. 2000 ; BEGER et al. 2001 ; KINSELLA et al. 2002 ). Among the advantages afforded by dominant genetics is its ability to define cellular components, e.g., proteins, which are vunerable to perturbation and the utility of dominant agents as probes to identify their binding targets within cells (GUDKOV et al. 1994 ; CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ).
Here, we use a combination of (i) a fluorescent reporter linked to a promoter; (ii) an optimized clonal cell line engineered with the reporter; and (iii) a flow sorter to isolate, from an expression library, dominant modulators that induce or suppress the expression of the reporter. We investigated the retinoic acid pathway that regulates the expression of genes, many of which contain a retinoic-acid-responsive element (DE LUCA 1991 ; CHAMBON 1996 ). Retinoids regulate gene expression via the RAR-RXR nuclear receptors that act in conjunction with ancillary corepressors and coactivators (reviewed in FREEDMAN 1999 ; XU et al. 1999 ). Retinoids can induce differentiation, growth inhibition, and/or apoptosis in certain cell types and are used to treat specific cancers of the blood and skin diseases (reviewed in SMITH et al. 1992 ; JOHNSON and CHANDRARATNA 1999 ; SAURAT 1999 ; DRAGNEV et al. 2000 ; HANSEN et al. 2000 ; FENAUX et al. 2001 ). Using a retinoic-acid-responsive element-containing promoter to drive a reporter in a human cell line, we identified dominant agents that either induced or suppressed reporter expression.
MATERIALS AND METHODS.mtp, 百拇医药
Generation of expression constructs:.mtp, 百拇医药
A dominant-negative RAR{alpha} receptor, RAR{alpha} {Delta} 403 (TSAI et al. 1992 ), lacking 59 amino acid residues of the C terminus of the wild-type protein, was PCR amplified from a human brain cDNA library. This cDNA fragment was subcloned into a retroviral vector, pVT352, as an in-frame C-terminal fusion to a nonfluorescent variant of green fluorescent protein (dGFP; Tyr-to-Phe mutation at codon 66 in the chromophore). This fusion was placed under the regulation of a constitutive cytomegalovirus promoter; the pVT352 vector carried a neomycin marker for selection of stably transduced cells. One of these scaffolded RAR{alpha} {Delta} 403 clones was verified by DNA sequencing and used in subsequent experiments..mtp, 百拇医药
A random-primed cDNA library of 12 x 106 clones was generated from human placental tissue. Poly(A) mRNA was isolated from placental tissue using standard protocols. The mRNA was used to generate the first strand of cDNA using random nonamer primers linked to a SfiI restriction site (SfiI-A: 5'-GGCCGAGGCGGCC-3'). Second-strand synthesis was accomplished using random hexamer primers linked to a second SfiI restriction site (SfiI-B: 5'-GGCCATAATGGCC-3'). The resulting material was then PCR amplified, size selected, and directionally cloned into compatible SfiI-B to SfiI-A sites of pVT352, where the cDNA fragments were expressed as C-terminal fusions to the nonfluorescent dGFP scaffold.
To make out-of-frame versions of given dominant agents, cDNA fragments were cloned into a variant pVT352 vector that has an altered 3' fusion junction to the dGFP scaffold such that the cDNA sequences present in a different reading frame. All specific expression constructs were subsequently transformed into DH10B Escherichia coli (Life Technologies) and clones were independently sequenced to confirm that desired constructs were generated and that mutations were not introduced. DNAs harvested from sequenced verified clones were then packaged in 293gp cells and the retroviruses produced were used to transduce cells.op!+^gy, 百拇医药
To perform epistasis studies, we replaced the neomycin resistance gene contained in pVT352 with a hygromycin gene from pCMVLacI (Stratagene, La Jolla, CA). The 249-bp R3 sequence was cloned into the SfiI sites downstream of the dGFP and sequence verified. Transduced cells were selected for 8 days in 200 µg/ml hygromycin (Life Technologies).op!+^gy, 百拇医药
Retroviral packaging and transduction:
Infectious retroviral particles were generated by transiently transfecting retroviral plasmid (Moloney murine leukemia virus-based) and VSV-G expression plasmid using LipofectAmine (Life Technologies) into 293gp cells, which stably expressed gag and pol gene products (provided by Inder Verma, Salk Institute). Media from transfected cells was collected after 72 hr and filtered through a 0.45-µm membrane. Transfected cells were grown in Dulbecco's modified Eagle's medium, 10% fetal bovine serum (FBS), and 2 mM L-glutamine (Life Technologies) at 37° in a humidified incubator.ym[89, http://www.100md.com
Transduction of WM35 cells was performed by addition of 20% vol/vol of retroviral supernatant to media and 8 µg/ml polybrene (Sigma, St. Louis) to the media. Cells were exposed to the retrovirus for 24 hr, followed by media change and a 24-hr recovery period prior to selection of retroviral linked selection markers (neomycin, puromycin, or hygromycin). A 20% retroviral supernatant typically resulted in transduction of 30–60% of these human melanoma cells; in some instances, 2% transductions were performed to ensure a single- or low-copy insertion.
WM35 and C8 cells:|{v-)38, http://www.100md.com
Human WM35 melanoma cells were a kind gift from Meenhard Herlyn (Wistar Institute). The WM35 cells were cultured in 80% keratinocyte basal medium and 20% L15 media (Clonetics, San Diego) supplemented with 2% FBS, 2 mM L-glutamine (Life Technologies). The C8 reporter line was generated by fluorescence-activated cell sorter (FACS) and clonal isolations from WM35 cells stably transduced with a retinoic-acid-regulated enhanced GFP reporter (RICHARDS et al. 1999 ). C8 cells were continually grown in media with 2 µg/ml puromycin (Sigma) to maintain the integrated reporter construct.|{v-)38, http://www.100md.com
C8 bioassays:|{v-)38, http://www.100md.com
Two types of bioassays were used to determine the activity of a given dominant agent: bioassay I, which identified modulators that repressed the C8 reporter and thereby made the cells GFP- (dim), and bioassay II, which identified modulators that induced the C8 reporter and thereby made the cells GFP+ (bright).
In bioassay I, C8 cells were transduced with retroviral expression constructs bearing a neomycin marker to introduce desired dominant agents, e.g., RAR{alpha} {Delta} 403. After ">="|%l0, http://www.100md.com
8 days of puromycin and G418 (550 µg/ml) selection, 5 x 105 of these cells were seeded in media with 2% FBS in T75 flasks. Five days later, the cells were trypsinized and analyzed for GFP expression by a Coulter (Hialeah, FL) Epics XL-MCL flow cytometer. Under these conditions, generally 80–95% of the FBS-treated C8 population were GFP+. The GFP- and GFP+ gates were based on the autofluorescence level of WM35 cells, whereby <1% of these naïve cells were present in the GFP+ gate.|%l0, http://www.100md.com
In bioassay II, 4.3 x 105 transduced and stably selected C8 cells were seeded in T75 flasks in media supplemented with 2% charcoal-stripped FBS serum (CBI, Cocalico Biologicals), which minimizes the presence of retinoids. After 5 days, the cells were then trypsinized and analyzed by flow cytometry for GFP expression. The GFP+ gate was set whereby generally <1% of the population from vector-control-transduced C8 cells were present in the gate. For the dose-response studies, C8 cells were maintained in media with 2% CBI for 4 days prior to the addition of given amounts of all-trans retinoic acid (ATRA; Sigma) dissolved in 100% ethanol. The cells were analyzed by flow cytometry 2 days after induction by ATRA.
Viability of cells in all bioassays was determined by propidium iodide staining. All analysis was performed on a single-cell viable subpopulation based on the flow cytometry of forward- and side-scatterplots (RICHARDS et al. 1999 ). In all data shown, the amount of cell death was generally <7%. The GFP signal was detected using an FL1 photomultiplier (525 ± 10 nm); voltages that clearly discriminated between the WM35 autofluorescence signal and the GFP+ signal were set.4lx}d[&, http://www.100md.com
Cycling of libraries:4lx}d[&, http://www.100md.com
Throughout the screens, desired GFP+ or GFP- cells were isolated by using a Coulter EPICS Elite fluorescence-activated cell sorter. The number of C8 cells transduced with retroviral cDNA libraries at each cycle allowed for more than fivefold coverage of the libraries. Genomic DNAs from WM35 cells of interest were isolated using a DNA kit from QIAGEN (Chatsworth, CA). Recombinant cDNA sequences introduced within the retroviral expression vector were recovered by PCR using oligonucleotides 5'-GGATCACTCTCGGCATGGACGAG-3' and 5'-ATTTTATCGAT GTTAGCTTGGCCATT-3', which flanked the cDNA fragment. In general, PCR was performed for 23 cycles under the following conditions: 94° for 30 sec, 68° for 2.5 min. The amplification reactions contained 10 mM KCl, 20 mM Tris-HCL, pH 8.8, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Triton X-100, 0.25 mM dNTPs, 1 µM primers, 1 µg of genomic DNA, and 0.1 units/µl Vent DNA polymerase (New England Biolabs, Beverly, MA). PCR products were digested with SfiI, directionally recloned into the pVT352 retroviral vector, and transformed into DH10B E. coli. DNA was then isolated and packaged in 293gp cells. Naïve C8 cells were transduced with these retroviruses and the neomycin and puromycin selected cells were reexamined in bioassay I or II.
Immunoprecipitation and Western analyses:a/8;}, 百拇医药
Immunoprecipitation experiments were performed by transiently cotransfecting PAT1 and R3 expression constructs into HEK293 cells by LipofectAmine using the manufacturer's protocol (Life Technologies). Two days post-transfection cells were rinsed in PBS, counted, and lysed in immunoprecipitation (IP) buffer: 25 mM HEPES (pH 7.5), 5 mM EDTA, 5 mM EGTA, 50 mM NaCl, 50 mM NaF, 10% (vol/vol) glycerol, 1% (vol/vol) Triton X-100, 2 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (Sigma Aldrich). The lysate was precleared followed by addition of 1 µg of anti-GFP monoclonal antibody (Roche Molecular Biochemicals) and incubation at 4° for 2 hr. Protein G plus Sepharose beads (Amersham Pharmacia Biotech) were subsequently added for 2 hr, followed by a wash with IP buffer. Agarose-protein complexes were resuspended in sample loading buffer processed for silver staining (Invitrogen, San Diego) and PAT1 Western blotting as previously described (ZHENG et al. 1998 ) using anti-PAT1 antibodies generously provided by Dr. Sanjay Pimplikar (Case Western Reserve University).
Immunoblots were performed on WM35 cells infected with each cDNA retroviral supernatant. The parental WM35 cell line was chosen because it does not express any GFP so fusions could be detected unambiguously using anti-GFP antibody. Western blots were carried out using standard protocols (HARLOW and LANE 1988 ). Briefly, total cell lysate from ~ 2 x 105 cells was loaded per well and separated at 200 V on a 10% Tris-Bis gel (Novex, San Diego). The proteins were then transferred to polyvinylidene fluoride membrane for 3 hr at 30 V. A monoclonal GFP antibody (Zymed, S. San Francisco) diluted 1/1000 was used for 1 hr at room temperature followed by a goat anti-mouse HRP-conjugated antibody diluted 1/5000. Proteins were indirectly detected using an enhanced chemiluminescence kit (Amersham).3p, 百拇医药
Gene expression profiling:3p, 百拇医药
Total RNA and poly(A) RNA was isolated from ~ 50 x 106 cells according to the protocols of QIAGEN. Messenger RNAs from two different C8 populations, test (R3 expressing or ATRA treated) and reference (R3of-expressing cells maintained in CBI) were labeled with either CY3 or CY5 and then competitively hybridized to a UniGEM V5 human microarray by Incyte Genomics. Detection and analysis of the ratio of transcript signals observed for the CY3 vs. CY5 probes were performed by Incyte. The calculated balanced differential expression (BDE) for each gene represents a ratio of test and reference signals for the microarray experiment. Genes that were induced (test signal greater than reference signal) were given a positive BDE number (+ test/reference), whereas repressed genes (test signal less than reference signal) were given a negative BDE number (- reference/test).
Yeast two-hybrid assays:^w, 百拇医药
To identify potential R3 interactors, a Lex-dGFP-R3 bait was used in a two-hybrid screen against three commercial libraries: Proquest fetal brain library (Invitrogen), Proquest HeLa library (Invitrogen), and Lex fetal brain library (BD Biosciences Clontech). yVT87 (EGY48) [MAT{alpha} , ura3, his3, trp1, LexAop(X6)-LEU2] yeast cells were transformed with pVT2527 that expresses dGFP-R3 fused to the LexA DNA-binding domain (LexA-dGFP-R3). The resulting strain was mated overnight in rich media (YEPD) to yVT99 [MATa, ura3, his3, trp1, LexAop(X6)-LEU2, LexAop(X8)-URA3] yeast harboring the two-hybrid cDNA libraries. The diploids were first selected on diploid selection media, synthetic-glucose plates lacking histidine and tryptophan, and then screened on interaction selection media, synthetic-glucose plates lacking histidine, tryptophan, uracil, and leucine for the Proquest libraries and on synthetic-galactose plates lacking histidine, tryptophan, uracil, and leucine for the Lex fetal brain library. The library coverage was at least threefold for each screen.
The validity of the yeast two-hybrid interaction was confirmed by a retest whereby plasmid DNA was isolated from colonies that emerged on the selection plates, propagated in E. coli, and retransformed into yVT99 yeast. These cells were then mated to yVT87 yeast harboring LexA-dGFP-R3, LexA-dGFP, or LexA-R3. The diploids were grown on synthetic-glucose plates lacking histidine and tryptophan and replica plated onto diploid selection media and interaction selection media. Increased growth on interaction selection media with LexA-dGFP-R3 compared to LexA-dGFP suggested that the interaction involved the R3 peptide. Several clones (e.g., BIP GenBank accession no. ; SFRS2, NM_003016) displayed interaction with the LexA-dGFP control, indicating that their readout did not require the presence of the R3 peptide; these define a class of false positives that associated with the bait scaffold. Another class of false positives was defined by "sticky" proteins that are commonly isolated from yeast two-hybrid screens, e.g., proteasome subunits and heat-shock proteins (see ).
RESULTS60y@f, 百拇医药
A retinoic-acid-responsive reporter line:60y@f, 百拇医药
Previously, we reported generation of a clonal human cell line, named C8, stably transduced with a retinoid-responsive-enhanced GFP reporter (RICHARDS et al. 1999 ). As shown in , when maintained in media supplemented with CBI (charcoal-stripped FBS to remove retinoids), 1–5% of the C8 population were present in the GFP+ gate; in contrast, 80–95% of these cells induced GFP expression (GFP+) upon treatment with 2% FBS. To further characterize the behavior of this retinoic-acid-regulated reporter, we performed a dose-response study in which C8 cells were treated with concentrations of ATRA ranging from 50 pM to 5 µM . The fluorescence histograms for this series of ATRA doses were essentially bimodal, producing a GFP- (dim) peak and a GFP+ (bright) peak. As the dose of ATRA increased, a greater fraction of induced GFP+ cells was detected while the fluorescence X-means remained relatively similar. The dose-response plot revealed that maximal induction occurred at >500 nM ATRA while the lower limit of reporter activation was detected at ~ 50 pM. Treatment with 5 pM concentrations of ATRA did not induce the reporter (data not shown). The bimodal response of the C8 reporter in human melanoma cells demonstrates that the cis- and trans-regulatory elements involved in this retinoic-acid-responsive system convert a graded ligand input into an all-or-none transcriptional readout on a single cell level. Similar analog-to-digital conversions (or threshold responses) have previously been described for other mammalian promoters (FIERING et al. 1990 ; KARTTUNEN and SHASTRI 1991 ; ROSSI et al. 2000 ).
fig.ommitted^istrw], http://www.100md.com
Figure 1. Isolation of a retinoid/serum-responsive clone. (A) Fluorescence histograms of C8 cell populations showing the number of cells on the y-axis and relative fluorescence units (RFU) on the x-axis. The C8 populations transduced with control pVT352 retroviral plasmid were maintained for 5 days in media supplemented with FBS (green) or CBI (red). Also shown are C8 cells transduced with the dominant-negative RAR403 construct that were subsequently maintained in FBS for 5 days before analysis by flow cytometry (blue). (B) A phase-contrast field of RAR403 transduced C8 cells maintained in media with FBS. (C) The same field demonstrating a speckled phenotype of GFP+ and GFP- cells as observed by fluorescence microscopy. The arrowheads indicate some cells that do not express GFP.^istrw], http://www.100md.com
fig.ommitted^istrw], http://www.100md.com
Figure 2. All-or-none response of the C8 reporter. Shown are the fluorescence histograms of C8 cells. The cells were initially maintained in media with CBI for 3 days, followed by treatment with the indicated concentrations of ATRA and flow cytometry analyses 2 days later. In the CBI-only control, no ATRA was added. Induction of the retinoic-acid-regulated GFP construct in these cells produces a bimodal distribution consisting of a GFP- population and a GFP+ population. Concentrations of ATRA >5 µM did not increase the fraction of cells in the GFP+ gate beyond that observed for the 5 µM ATRA sample (data not shown). The higher percentage of GFP+ cells in the FBS vs. the ">="
5 µM ATRA populations suggests that other factors besides retinoids can contribute to the activation of the C8 reporter. For comparison, the autofluorescence of parental WM35 melanoma cells lacking the reporter construct is presented in the top box. The GFP- gate was based on the autofluorescence level of naïve WM35 cells in which <1% of the population was present in the GFP+ gate.54nrul, 百拇医药
Modulation by a dominant-negative RAR{alpha} mutant:54nrul, 百拇医药
To validate the C8 reporter, we expressed RAR403, a known dominant-negative inhibitor of the RAR-RXR transcription complex, in C8 cells (TSAI et al. 1992 ; ISOGAI et al. 1997 ). We fused RAR403 cDNA to a nonfluorescent GFP scaffold (dGFP), chosen for its ability to enhance stability of fused peptides (ABEDI et al. 1998 ; SANDROCK et al. 2001 ), and expressed the fusion using a constitutive cytomegalovirus promoter in a retroviral vector. In contrast to parental C8 cells, C8 derivatives stably transduced with RAR403 retrovirus exhibited a bimodal distribution with 35% of the cells in the GFP- gate after 5 days of growth in media with FBS . This bimodal response likely reflected the penetrance of RAR403 in C8 cells because individual clones of RAR403-expressing cells had the same bimodal behavior (data not shown). When examined by fluorescence microscopy, single C8:RAR403 colonies displayed a "speckled" phenotype due to the mixture of GFP+ and GFP- cells in the colony . The repressor effect of RAR403 was consistent with its proposed mutant activity; i.e., it dimerizes with native RXRs and associates with retinoic-acid-responsive elements and corepressors to shut down transcription in a dominant fashion (TSAI et al. 1992 ). These data demonstrated that the C8 reporter responds as expected to known modulators of the RAR-RXR transcription machinery.
Repressors of C8 reporter expression:%km-jo%, 百拇医药
Using C8 cells, we searched for dominant agents that modulated expression of the retinoic-acid-responsive reporter. Two screens were performed: one to identify agents that repressed the C8 reporter and made the cells GFP-, or dim (class I), and one to identify agents that induced the C8 reporter and made the cells GFP+, or bright (class II).%km-jo%, 百拇医药
To find class I modulators, we expressed a human placental cDNA library fused to a nonfluorescent dGFP scaffold in C8 cells. The transduced C8 cells were maintained in media with FBS for 5 days (bioassay I) and the dimmest 5% of the F0 population was collected by FACS . The F1 subpopulation was expanded and subjected to two further cycles of dim sorting after which ~ 75% of cells were present in the GFP- gate , F3 subpopulation). Because colonies of C8 cells transduced with RAR403 exhibited a speckled appearance, we screened F3 cells for a similar phenotype. Colonies originating from single F3 cells were maintained in FBS for 5 days. When examined by fluorescence microscopy, ~ 10% of the F3 colonies were essentially GFP+ while ~ 40% of the colonies were GFP-. The fully bright colonies were presumably derived from cells with an intact retinoid/serum-response pathway. On the basis of control experiments, most of the completely dim colonies likely originated from cells with inactivating mutations in the retinoic-acid-regulated GFP retroviral reporter or with lesions at loci required to transactivate the reporter (data not shown).
fig.ommitted11x&, 百拇医药
Figure 3. Class I modulators that repress the C8 reporter. (A) Histogram profiles of C8 cells grown in 2% FBS for library sorts one through three. Approximately 25 x 106 C8 cells were transduced with the placental cDNA library retrovirus and selected for 8 days in G418-containing media. To isolate dominant agents that repressed the reporter, the cells were transferred into media with 2% FBS and, after 5 days, cells in the GFP- gate were recovered by FACS. These F1 GFP- sorted C8 cells were expanded and similarly sorted two more times to generate the F2 and F3 subpopulations. (B) C8 cells were transduced with the following constructs: vector, RAR403, and the in-frame (solid) and out-of-frame (of; stippled) versions of F791, F802, and F820; these were examined in bioassay I. (C) Flow cytometry data for vector (control) and the F820 in-frame and out-of-frame (F820of) constructs shown as representative data for the class I modulators. (D) Western analysis of WM35 cells transduced with each construct to confirm protein expression. With the exception of the RAR403 clone, each of the constructs expressed approximately similar protein levels and exhibited the expected size band based on the number of amino acid residues fused to the GFP scaffold (see ). The RAR403 clone was expressed at the mRNA level as detected by RT-PCR (data not shown). A nonspecific 60-kD band, recognized in all samples, serves as a control for equivalent loading.
Complementary DNAs from the F3 sublibrary were recovered from individual speckled colonies and a group of 10 randomly selected inserts were analyzed further. Relative to the vector control, 3 of the 10 clones (F797, F802, and F820) caused >36% of cells to fall within the GFP- gate . For comparison, 41% of dominant-negative RAR403-transduced C8 cells were in the GFP- gate. The bimodal fluorescence histogram of C8 cells containing F820 is shown as representative data .s'h-hun, 百拇医药
To determine if the three class I modulators acted at the protein or RNA level, each of the three sequences was expressed in an alternate reading frame (out of frame, of;. In the case of clone F797, the in-frame and out-of-frame constructs displayed a similar increase in cells in the GFP- gate, suggesting RNA-level activity. In contrast, the out-of-frame constructs of F802 and F820 did not produce a significant elevation in the percentage of GFP- cells above the vector control. To confirm that these out-of-frame constructs produce protein (and by inference, mRNA), we performed Western blots with an anti-GFP antibody. Each of the constructs expressed approximately equivalent amounts of protein at the expected size. Taken together, these results suggested that F802 and F820 act at the protein level. Of interest, expression of RAR403 was not detectable on this immunoblot even though C8 reporter repression was observed in cells transduced with this construct . This result could be explained by the observation that a high level of expression of the dominant-negative RAR{alpha} polypeptide is cytotoxic in these human cells (data not shown).
The identities of the dominant agents were investigated by sequence comparisons with public nucleotide databases. Matches were identified for F797 and F802, but not F820 (. The F797 clone consisted of 572 nucleotides from the sense strand of the HSP90 mRNA 5' untranslated region. Clone F802 encoded 64 amino acid residues from the central portion of cyclophilin B, including the peptidyl-prolyl cis-trans isomerase domain, fused in-frame with the dGFP scaffold. The F820 sequence consisted of a 249-base insert that encoded a 29-amino-acid peptide fused to dGFP.w(:-0k, 百拇医药
fig.ommittedw(:-0k, 百拇医药
Table 1. Dominant modulators of retinoid signalingw(:-0k, 百拇医药
Inducers of C8 reporter expression:w(:-0k, 百拇医药
In addition to identifying class I modulators, we screened for class II modulators that induced the C8 reporter and thereby produced GFP+ cells. C8 cells harboring the expression library were maintained in media with CBI for 5 days (bioassay II), after which the brightest 0.2% of the population was collected by FACS. This subpopulation was grown and genomic DNA was isolated from the cells. The retroviral inserts were recovered by PCR and amplified fragments were recloned into retroviral expression vector, packaged, and used to transduce naïve C8 cells. After neomycin selection, the GFP+ subpopulation was again recovered by FACS from cells maintained in CBI. This cycling process was repeated five more times. By the sixth cycle, the fraction of cells in the GFP+ gate was 23-fold higher than that of control cells .
fig.ommitted[qk?'x, 百拇医药
Figure 4. Class II modulators that induce the C8 reporter. (A) A comparison of the percentage of GFP+ cells for library- and vector-control-transduced C8 cells. In the first step, ~ 30 x 106 C8 cells were transduced with the placental cDNA library retrovirus and selected for 8 days in G418-containing media. To isolate dominant agents that induced the retinoic-acid-regulated GFP reporter, the cells were transferred into media with 2% CBI and, after 5 days, cells in the GFP+ gate were recovered by FACS. Retroviral cDNA inserts from the sorted cells were recovered by PCR, recloned into retroviral vector, packaged, and used to transduce naïve C8 cells. The GFP+ cells in this sublibrary-transduced C8 population were isolated by FACS. (B) A histogram of cycle-6-transduced C8 cells compared to vector-control-transduced cells. (C) C8 cells were transduced with the following constructs: control vector and the in-frame (solid) and out-of-frame (of; stippled) versions of R1, R2, R3, and R44, which were examined in bioassay II. (D) Flow cytometry data for vector and R3 in-frame and out-of-frame (R3of) constructs shown as representative data for the class II modulators. (E) Western analysis of WM35 cells transduced with each construct to confirm protein expression. The R1of clone was expressed at the mRNA level as detected by reverse-transcription-based PCR (data not shown).
We sequenced 134 clones from the cycle 6 sublibrary to ascertain identities and representation. The most abundant clone, R3, comprised 37% of sequences analyzed, while the next most abundant clones, R1 and R2, accounted for 13 and 9% of the sequences, respectively. We tested the 15 most frequent clones independently in C8 bioassay II and found that 4 of them exhibited significant signals above the vector control. The strongest effect was elicited by the R3 clone: 54% of the R3-infected C8 cells were present in the GFP+ gate compared to 1% in control cells . Sequence analysis revealed that the R1, R2, R3, and R44 clones encoded peptides of nine residues or fewer fused to the dGFP scaffold . The possibility that the peptide fusions of these clones to the dGFP scaffold made the mutant protein fluorescent was eliminated by showing that expression of these clones in WM35 or other cells did not produce any detectable fluorescence signal by flow cytometry (data not shown). By comparing the effects of in-frame and out-of-frame constructs and by performing Western analyses for expression , all of the class II modulators were shown to act as peptides. The out-of-frame R1 construct, although it produced significantly less protein compared to R1, expressed similar levels of mRNA as determined by reverse-transcription-based PCR (data not shown).
Mechanism of action of the R3 inducer:ziqy, 百拇医药
We performed genetic epistasis studies to examine the relationship between R3, the strongest class II modulator, and RAR403, a known dominant-negative factor of the retinoid pathway. Using bioassay I as the readout, we found that a lower percentage of GFP- cells was present in the population expressing both RAR403 and R3 compared to cells expressing RAR403 alone (11 vs. 28%, ). Thus, R3 interfered with the ability of RAR403 to block pathway induction. Using bioassay II as the readout, fewer bright cells were present in the population expressing both proteins than in cells expressing only R3 (22 vs. 35%). Thus, RAR403 partially blocked pathway induction by R3. Taken together, these data indicated that each factor (RAR403 and R3) antagonizes the activity of the other.ziqy, 百拇医药
fig.ommittedziqy, 百拇医药
Figure 5. Genetic epistasis between R3 and dominant-negative RAR{alpha} . C8 cells were initially transduced with either dGFP-R3 fusion or control dGFP (Vector) expressed from a constitutive cytomegalovirus promoter in a retroviral vector carrying a hygromycin selection marker. After selection with hygromycin, some of the cells were secondarily transduced with a retrovirus bearing dGFP-RAR403 and subsequently selected with G418 to generate a population carrying both constructs. These constructs were expressed from a constitutive cytomegalovirus promoter. Shown on the left (solid bars) are the percentages of the GFP- population of C8 cells, transduced with the indicated constructs and subjected to bioassay I (i.e., seeded, maintained in media with 2% FBS, and then analyzed by flow cytometry after 5 days). Shown on the right (hatched) are the percentages of the GFP+ population of C8 cells subjected to bioassay II (i.e., seeded, maintained in media supplemented with 2% CBI, and then analyzed after 6 days). Averaged data for three replicate samples, along with standard deviations, are shown for each construct.
A second approach, gene expression profiling by microarray analyses, was used to compare the effects of R3 on C8 cells to that elicited by ATRA. The primary purpose of these studies was to determine if R3 regulated the expression of a similar set of genes as retinoids. Transcript levels for ~ 9000 genes were compared in C8 cells expressing dGFP-R3 or C8 cells expressing dGFP-R3of that were treated with ATRA. In the case of the R3-expressing cells, there were only 30 genes (outliers) whose expression was up- or downregulated >1.7-fold (; ). These data suggest that R3 causes a limited change in the global expression profile of these cells. In contrast, ATRA treatment of C8 cells expressing R3of caused a more dramatic shift in transcript profile, with 285 genes altered in expression level >1.7-fold (). Notably, the set of 30 genes (>1.7-fold modulated) in R3-expressing cells was essentially a subset of the ATRA-treated data. Overall, 29 of the 30 outliers in the R3 data set responded similarly in the ATRA experiment (+ for induced, - for repressed; ). Moreover, 17 of the 30 outliers in the R3 data set were also outliers of the same sign in the ATRA experiment Covariance calculations revealed that this correlation between the R3 subgroup and the equivalent genes in ATRA-treated cells was highly significant (covariance = 1.89, P << 0.001; ).
fig.ommitted@, http://www.100md.com
Figure 6. Transcript microarray analysis of R3. Scatter plot showing normalized expression levels of 7529 genes (genes not expressed in both samples have been removed) in the following types of C8 cells. (A) Cells expressing dGFP-R3 or dGFP-R3of maintained in CBI. (B) dGFP-R3of cells treated with CBI + ATRA or maintained in CBI alone. Red dots correspond to the subset of outliers of R3 vs. R3of (for which the R3/R3of balanced differential expression is > +1.7 or < -1.7) that are also outliers, with the same sign, in the CBI + ATRA vs. CBI data. (C) Histogram of covariances calculated using the 30-member outlier subgroup for dGFP-R3-expressing cells compared to the same group of genes in ATRA-treated cells (arrow) or to sets of 30 randomly selected genes in the ATRA data set (distribution). Covariance was calculated according to the formula cov(x, y) = (1/N) (xi -)(yi - ), where N is the number of data points in the set, x is the fractional increase (positive) or decrease (negative) in the expression of a particular gene in the R3 vs. R3of data set, and y is the fractional change in the CBI + ATRA vs. CBI data set. The covariance of the 30-member outlier set was 1.89. The mean covariance of this set computed with 1000 random 30-member sets was 0.003 and the standard deviation was 0.093.
fig.ommittedn9xmar, 百拇医药
Table 2. Outlier genes subgroup of R3 vs. R3of microarray experimentn9xmar, 百拇医药
R3 interactors:n9xmar, 百拇医药
To identify proteins that associate with R3, we executed a yeast two-hybrid screen. dGFP-R3 was fused downstream of a LexA DNA-binding domain and used as a bait to screen activation domain-cDNA libraries derived from fetal human brain tissue and cervical HeLa cells. Briefly, 400 putative positives were isolated from the initial screens. Sequence analyses revealed that 195 of these clones were out of frame or encoded known yeast two-hybrid false positives (see MATERIALS AND METHODS). A total of 205 clones were retested with LexA-dGFP-R3 or LexA-dGFP to verify specificity of the interaction with R3. Nineteen of the clones passed the retest by displaying stronger interaction with the LexA-dGFP-R3 bait compared to the LexA-dGFP control. These 19 positives consisted of partial or full-length in-frame cDNAs encoding four proteins: PAT1 (GenBank accession no. ), HIRIP4 (XM_040816), TCTEL-1 (), and {alpha} -B-crystallin (NM_001885). Specifically, two of the four cDNA clones, TCTEL-1 and {alpha} -B-crystallin, associated weakly with LexA-GFP; importantly, PAT1 and HIRIP did not. In addition, we executed a second study whereby we generated a LexA-R3 bait and used it to ascertain whether any of the four cDNAs interacted with the R3 sequences in a scaffold (GFP)-independent manner. Of the four cDNA clones, only PAT1 displayed interaction with LexA-R3 in the assay , indicating that the nine amino acid residues encoded by R3 were sufficient for enabling interaction with the protein.
fig.ommitted44!, 百拇医药
Figure 7. Interaction of R3 with PAT1. (A) Yeast two-hybrid auxotrophy readout of the indicated baits on a LexA DNA-binding domain with a prey of PAT1 fused to the C terminus of an activation domain (AD). To demonstrate that dGFP-R3 binds to PAT1 in human cells, the two proteins were exogenously co-expressed in HEK293 cells and the resulting soluble fraction of cell extracts was subjected to immunoprecipitation and protein analyses. Data for two independent PAT1 clones, no. 5 and no. 10, are shown. (B) The out-of-frame version of R3, dGFP-R3of, was also analyzed as a control for the specificity of the interaction between GFP-R3 and PAT1. Samples: lane 1, dGFP-R3 and PAT1 clone no. 5; lane 2, dGFP-R3of and PAT1 clone no. 5; lane 3, dGFP-R3 and PAT1 clone no. 10; and lane 4, dGFP-R3of and PAT1 clone no. 10. Specifically, IP was performed using anti-GFP antibodies and the precipitate was examined by (B) silver-stain detection of proteins fractionated by polyacrylamide gel electrophoresis (PAGE) or (C) by Western blotting using anti-PAT1 antibodies. (D) A control Western blot was performed to show equivalent PAT1 expression in the different samples.
Examination of the transcript microarray data from C8 cells revealed that PAT1 was expressed in these human melanoma cells. Specifically, the microarray analyses quantitated signals of 601 arbitrary expression units, 716 units and 595 units in C8 cells treated with ATRA, R3, or R3of, respectively, where <250 units was considered no expression. The microarray data did not show any significant induction or repression (> or <1.7-fold) of PAT1 in the retinoic-acid-induced or R3-expressing C8 cells relative to the R3of-expressing cells.'v[k, 百拇医药
We investigated whether dGFP-R3 and PAT1 interacted in mammalian cells. We initially examined C8 and WM35 cells but found that the expression levels of endogenous PAT1 protein in these melanoma cells were not detectable by Western analyses using anti-PAT1 antibodies; moreover, the cells were not amenable to transfections. Full-length PAT1 was therefore transfected into human HEK293 cells along with either dGFP-R3 or dGFP-R3of. Subsequently, cell extracts were prepared and subjected to IP-Western analyses using anti-GFP antibodies for IP and anti-PAT1 antibodies for the Western blots. As shown in , association between PAT1 and dGFP-R3 was observed in two independent experiments (lanes 1 and 3). In contrast, the out-of-frame R3 construct (lanes 2 and 4) was successfully immunoprecipitated by anti-GFP antibodies but failed to pull down PAT1 . As a control, we demonstrated that PAT1 was equivalently expressed in the different cell extracts
DISCUSSION9v, 百拇医药
Several protein scaffolds, including GFP, thioredoxin, and inactivated staphylococcal nuclease, have been previously used in screens for peptide aptamers that evoke dominant effects in cellular processes (COLAS et al. 1996 ; CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ). The primary reason for using a protein scaffold for the presentation of peptides is stability, as studies have shown that high levels of expression of a scaffolded peptide can be achieved within cells (SANDROCK et al. 2001 ). In addition, the scaffold serves as a tag for the dominant agent, e.g., for monitoring expression, subcellular localization, or purification. However, the use of a scaffold in this type of context raises other issues. For example, the scaffold is sizable (30 kD in the case of GFP) and may thus stearically hinder the association of peptide fusions to certain proteins. Although the scaffold is assumed to be "inert," it may not be so in certain instances. It is conceivable that the scaffold contributes to the association of the peptide fusion with its target. Moreover, once bound to a target protein, the scaffold could occlude the binding of other molecules that normally interact with the protein.
All of the dominant agents from our screen were fused to the C terminus of the dGFP scaffold. By comparing the activity and expression levels of in- and out-of-frame versions of the clones, we demonstrate that the activities of six of the seven dominant agents reported here reside in the expressed protein. Only the F797 clone appeared to act at the RNA level. Of the isolated clones, F802 was the sole one that encoded a fragment of a known protein (cyclophilin B). The other five peptide modulators consisted of peptides ranging from 3 to 29 amino acid residues. The low frequency of protein fragments obtained from the screen is primarily due to the fact that the random-primed cDNA expression library used in these selections is expected to contain a majority of clones that express peptides derived from dGFP fusions to out-of-frame coding or untranslated sequences.+./3n, 百拇医药
We chose to investigate the R3 peptide most thoroughly because of the strength of its effect, and because it mimics ATRA action and therefore has potential relevance to retinoid-based therapies such as promyelocytic leukemia treatment (SMITH et al. 1992 ; VARGHESE et al. 1999 ; FENAUX et al. 2001 ). Through genetic epistasis studies with a dominant-negative RAR{alpha} mutant , we demonstrated that R3 acts within the retinoid pathway. Analysis of the microarray data suggests that R3 activity is a subset of the biological effects of ATRA (). R3 appears to have striking specificity; only a single gene from the 30-member outlier group in R3 vs. R3of was negatively correlated with the corresponding gene in ATRA-treated cells (). Thus, the phenotypic consequences of R3 expression appear to be largely those that were selected using the retinoic-acid-responsive reporter, namely, induction of the pathway.
Many other dominant peptide aptamers that have been characterized in other systems have been shown to function as inhibitors of protein function (COLAS et al. 1996 ; CAPONIGRO et al. 1998 ; KOLONIN and FINLEY 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ; BLUM et al. 2000 ; BUTZ et al. 2000 ). It is therefore plausible that R3 inhibits the activity of a repressor to induce reporter expression in C8 cells. Other mechanistic explanations for R3 action were not supported by the data. For example, expression of R3 did not alter fluorescence in WM35 cells that express enhanced GFP from a constitutive SV40 promoter (data not shown). This result excludes the possibility that R3 stabilizes the GFP mRNA or protein.1t6%[!, 百拇医药
As previously shown, if a peptide acts on another protein to elicit its effect, protein-protein interaction assays such as the yeast two-hybrid system can be used to identify candidate targets (CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ). Using yeast two-hybrid assays, we showed that coding segments from PAT1, HIRIP4, TCTEL-1, and {alpha} -B-crystallin interacted with a Lex-dGFP-R3 bait. PAT1 and HIRIP4 did not interact with the Lex-dGFP control bait, indicating that the nine C-terminal amino acid residues encoded by R3 were necessary for interaction with these cDNA clones. Moreover, only PAT1 interacted with R3 in a dGFP-independent manner in the yeast two-hybrid assay and co-immunoprecipitation studies showed that PAT1 binds to dGFP-R3 in human cells . PAT1 was initially described by ZHENG et al. 1998 as a kinesin-like protein that associates with the basolateral sorting signal of amyloid precursor protein, as well as with microtubules and membranes. The primary sequence of PAT1 contains numerous motifs of interest, including four kinesin light-chain-like repeats, two leucine zipper domains, a "LXLL" motif (NR box) found in transcriptional coregulators, a PEST protein-degradation sequence, as well as several putative nuclear localization signals. GAO and PIMPLIKAR 2001 have reported that PAT1 degradation is augmented by the expression of certain APP proteolytic fragments, which also represses retinoic-acid-responsive gene expression. Intriguingly, this observation is conversely consistent with preliminary studies that reveal that GFP-R3, an inducer of the retinoid pathway, increases the levels of PAT1 protein in mammalian cells (S. PIMPLIKAR, personal communication). Further work is therefore warranted to elucidate any role that PAT1 may have in retinoid signaling.
It is less certain how the other dominant agents act. The F797 RNA-level modulator cannot function as a classic antisense molecule because it matches the sense strand of HSP90. It could behave as a cross-hybridizing antisense RNA or as an RNA aptamer (FIRE 1999 ; HERMANN and PATEL 2000 ). The three amino acid peptide modulators (R1, R2, and R44) are of special interest because of their size. These dominant agents may act at least partly through some residues of the dGFP scaffold. F802 consists of the central peptidyl-prolyl isomerase domain of cyclophilin B (CypB), a member of a family of enzymes that influences protein structure and function (PRICE et al. 1991 ). Native CypB mediates several cellular processes (SWANSON et al. 1992 ; RYCYZYN et al. 2000 ), possesses an N-terminal signal sequence, and has been reported to be secreted (ALLAIN et al. 1999 ). Since F802 appears to be localized within the cell, it seems unlikely that F802 functions in a classic dominant-negative manner, unless CypB has intracellular activities as well. It is conceivable that F802 interacts with one or more of the ligands that another cyclophilin binds and thereby interferes with its function.
CONCLUSIONS#ldm, 百拇医药
Using the optimized retinoic-acid-responsive C8 reporter line, we executed two en masse genetic screens that identified repressors and inducers of the reporter. Six of seven of these modulators functioned as peptides. A single clone comprised the coding fragment of a known gene, CypB. Initial transcript microarray and genetic epistasis studies on the mechanism of action of R3 strongly supported its role as a specific inducer of the retinoid pathway. PAT1 was identified as a protein that interacts with R3. The dominant agents isolated in this study should be useful as reagents to probe and further understand the underpinnings of retinoid signaling. Moreover, because retinoids are used for the treatment of certain leukemias and skin diseases, and because the R3 peptide appears to induce the pathway in a manner similar to ATRA, it will be of interest to determine if R3 or its target has therapeutic utility.#ldm, 百拇医药
ACKNOWLEDGMENTS
We are indebted to Dr. Nikunj Somia for sharing his expertise on retroviruses and Dr. Sanjay Pimplikar for information about PAT1 and for kindly providing antibodies against the protein. We are grateful to Anthony Gaglio, Sharon Malmstrom, Stephen Hansen, and Rex Malmstrom for excellent technical assistance. We wish to thank Drs. Giordi Caponigro, Michael Feldhaus, Michael Pierce, Mark Poritz, and Mahendra Rao for input throughout this project and for comments on the manuscript.tx, 百拇医药
Manuscript received October 4, 2002; Accepted for publication December 3, 2002.tx, 百拇医药
LITERATURE CITEDtx, 百拇医药
ABEDI, M. R., G. CAPONIGRO, and A. KAMB, 1998 Green fluorescent protein as a scaffold for intracellular presentation of peptides. Nucleic Acids Res. 26:623-630.tx, 百拇医药
ALLAIN, F., S. DURIEUX, A. DENYS, M. CARPENTIER, and G. SPIK, 1999 Cyclophilin B binding to platelets supports calcium-dependent adhesion to collagen. Blood 94:976-983.tx, 百拇医药
AXENOVICH, S. A., A. R. KAZAROV, A. D. BOIKO, G. ARMIN, and I. B. RONINSON et al., 1998 Altered expression of ubiquitous kinesin heavy chain results in resistance to etoposide and hypersensitivity to colchicine: mapping of the domain associated with drug response. Cancer Res. 58:3423-3428.
BEGER, C., L. N. PIERCE, M. KRUGER, E. G. MARCUSSON, and J. M. ROBBINS et al., 2001 Identification of Id4 as a regulator of BRCA1 expression by using a ribozyme-library-based inverse genomics approach. Proc. Natl. Acad. Sci. USA 98:130-135.t4-^s;^, 百拇医药
BLUM, J. H., S. L. DOVE, A. HOCHSCHILD, and J. J. MEKALANOS, 2000 Isolation of peptide aptamers that inhibit intracellular processes. Proc. Natl. Acad. Sci. USA 97:2241-2246.t4-^s;^, 百拇医药
BUTZ, K., C. DENK, A. ULLMANN, M. SCHEFFNER, and F. HOPPE-SEYLER, 2000 Induction of apoptosis in human papillomaviruspositive cancer cells by peptide aptamers targeting the viral E6 oncoprotein. Proc. Natl. Acad. Sci. USA 97:6693-6697.t4-^s;^, 百拇医药
CAPONIGRO, G., M. R. ABEDI, A. P. HURLBURT, A. MAXFIELD, and W. JUDD et al., 1998 Transdominant genetic analysis of a growth control pathway. Proc. Natl. Acad. Sci. USA 95:7508-7513.t4-^s;^, 百拇医药
CHAMBON, P., 1996 A decade of molecular biology of retinoic acid receptors. FASEB J. 10:940-954.t4-^s;^, 百拇医药
COLAS, P., B. COHEN, T. JESSEN, I. GRISHINA, and J. MCCOY et al., 1996 Genetic selection of peptide aptamers that recognize and inhibit cyclin-dependent kinase 2. Nature 380:548-550.
DE LUCA, L. M., 1991 Retinoids and their receptors in differentiation, embryogenesis, and neoplasia. FASEB J. 5:2924-2933.##n7$4j, 百拇医药
DRAGNEV, K. H., J. R. RIGAS, and E. DMITROVSKY, 2000 The retinoids and cancer prevention mechanisms. Oncologist 5:361-368.##n7$4j, 百拇医药
FENAUX, P., C. CHOMIENNE, and L. DEGOS, 2001 All-trans retinoic acid and chemotherapy in the treatment of acute promyelocytic leukemia. Semin. Hematol. 38:13-25.##n7$4j, 百拇医药
FIERING, S., J. P. NORTHROP, G. P. NOLAN, P. S. MATTILA, and G. R. CRABTREE et al., 1990 Single cell assay of a transcription factor reveals a threshold in transcription activated by signals emanating from the T-cell antigen receptor. Genes Dev. 4:1823-1834.##n7$4j, 百拇医药
FIRE, A., 1999 RNA-triggered gene silencing. Trends Genet. 15:358-363.##n7$4j, 百拇医药
FREEDMAN, L. P., 1999 Increasing the complexity of coactivation in nuclear receptor signaling. Cell 97:5-8.##n7$4j, 百拇医药
GAO, Y. and S. W. PIMPLIKAR, 2001 The gamma-secretase-cleaved C-terminal fragment of amyloid precursor protein mediates signaling to the nucleus. Proc. Natl. Acad. Sci. USA 98:14979-14984.
GEYER, C. R., A. COLMAN-LERNER, and R. BRENT, 1999 "Mutagenesis" by peptide aptamers identifies genetic network members and pathway connections. Proc. Natl. Acad. Sci. USA 96:8567-8572.in;$, 百拇医药
GUDKOV, A. V., A. R. KAZAROV, R. THIMMAPAYA, S. A. AXENOVICH, and I. A. MAZO et al., 1994 Cloning mammalian genes by expression selection of genetic suppressor elements: association of kinesin with drug resistance and cell immortalization. Proc. Natl. Acad. Sci. USA 91:3744-3748.in;$, 百拇医药
HANSEN, L. A., C. C. SIGMAN, F. ANDREOLA, S. A. ROSS, and G. J. KELLOFF et al., 2000 Retinoids in chemoprevention and differentiation therapy. Carcinogenesis 21:1271-1279.in;$, 百拇医药
HARLOW, E., and D. LANE, 1988 Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.in;$, 百拇医药
HERMANN, T. and D. J. PATEL, 2000 Adaptive recognition by nucleic acid aptamers. Science 287:820-825.in;$, 百拇医药
HERSKOWITZ, I., 1987 Functional inactivation of genes by dominant negative mutations. Nature 329:219-222.
HOLZMAYER, T. A., D. G. PESTOV, and I. B. RONINSON, 1992 Isolation of dominant negative mutants and inhibitory antisense RNA sequences by expression selection of random DNA fragments. Nucleic Acids Res. 20:711-717.-d#6e8, 百拇医药
ISOGAI, M., M. V. CHIANTORE, M. HAQUE, G. SCITA, and L. M. DE LUCA, 1997 Expression of a dominant-negative retinoic acid receptor construct reduces retinoic acid metabolism and retinoic acid-induced inhibition of NIH-3T3 cell growth. Cancer Res. 57:4460-4464.-d#6e8, 百拇医药
JOHNSON, A. and R. A. CHANDRARATNA, 1999 Novel retinoids with receptor selectivity and functional selectivity. Br. J. Dermatol. 140(Suppl. 54):12-17.-d#6e8, 百拇医药
KAMB, A. and D. H. TENG, 2000 Transdominant genetics, peptide inhibitors and drug targets. Curr. Opin. Mol. Ther. 2:662-669.-d#6e8, 百拇医药
KARTTUNEN, J. and N. SHASTRI, 1991 Measurement of ligand-induced activation in single viable T cells using the lacZ reporter gene. Proc. Natl. Acad. Sci. USA 88:3972-3976.-d#6e8, 百拇医药
KINSELLA, T. M., C. T. OHASHI, A. G. HARDER, G. C. YAM, and W. LI et al., 2002 Retrovirally delivered random cyclic peptide libraries yield inhibitors of interleukin-4 signaling in human B cells. J. Biol. Chem. 277:37512-37518.
KOLONIN, M. G. and R. L. FINLEY, JR., 1998 Targeting cyclin-dependent kinases in Drosophila with peptide aptamers. Proc. Natl. Acad. Sci. USA 95:14266-14271.79kwq], http://www.100md.com
LEVENSON, V. V., E. LAUSCH, D. J. KIRSCHLING, E. V. BROUDE, and I. A. DAVIDOVICH et al., 1999 A combination of genetic suppressor elements produces resistance to drugs inhibiting DNA replication. Somat. Cell Mol. Genet. 25:9-26.79kwq], http://www.100md.com
NORMAN, T. C., D. L. SMITH, P. K. SORGER, B. L. DREES, and S. M. O'ROURKE et al., 1999 Genetic selection of peptide inhibitors of biological pathways. Science 285:591-595.79kwq], http://www.100md.com
PRICE, E. R., L. D. ZYDOWSKY, M. J. JIN, C. H. BAKER, and F. D. MCKEON et al., 1991 Human cyclophilin B: a second cyclophilin gene encodes a peptidyl-prolyl isomerase with a signal sequence. Proc. Natl. Acad. Sci. USA 88:1903-1907.79kwq], http://www.100md.com
RICHARDS, B., J. KARPILOW, C. DUNN, L. ZHARKIKH, and A. MAXFIELD et al., 1999 Creation of a stable human reporter cell line suitable for FACS-based, transdominant genetic selection. Somat. Cell Mol. Genet. 25:191-205.
ROSSI, F. M., A. M. KRINGSTEIN, A. SPICHER, O. M. GUICHERIT, and H. M. BLAU, 2000 Transcriptional control: rheostat converted to on/off switch. Mol. Cell 6:723-728.p, http://www.100md.com
RYCYZYN, M. A., S. C. REILLY, K. O'MALLEY, and C. V. CLEVENGER, 2000 Role of cyclophilin B in prolactin signal transduction and nuclear retrotranslocation. Mol. Endocrinol. 14:1175-1186.p, http://www.100md.com
SANDROCK, T. M., B. RISLEY, B. T. RICHARDS, M. A. PORITZ, and H. A. AUSTIN et al., 2001 Exogenous peptide and protein expression levels using retroviral vectors in human cells. Mol. Ther. 4:398-406.p, http://www.100md.com
SAURAT, J. H., 1999 Retinoids and psoriasis: novel issues in retinoid pharmacology and implications for psoriasis treatment. J. Am. Acad. Dermatol. 41:S2-S6.p, http://www.100md.com
SMITH, M. A., D. R. PARKINSON, B. D. CHESON, and M. A. FRIEDMAN, 1992 Retinoids in cancer therapy. J. Clin. Oncol. 10:839-864.p, http://www.100md.com
SWANSON, S. K., T. BORN, L. D. ZYDOWSKY, H. CHO, and H. Y. CHANG et al., 1992 Cyclosporin-mediated inhibition of bovine calcineurin by cyclophilins A and B. Proc. Natl. Acad. Sci. USA 89:3741-3745.
TSAI, S., S. BARTELMEZ, R. HEYMAN, K. DAMM, and R. EVANS et al., 1992 A mutated retinoic acid receptor-alpha exhibiting dominant-negative activity alters the lineage development of a multipotent hematopoietic cell line. Genes Dev. 6:2258-2269.??:}%, 百拇医药
VARGHESE, L., A. JANCKILA, and L. T. YAM, 1999 Acute promyelocytic leukemia: new methods in diagnosis and treatment. J. Ky. Med. Assoc. 97:61-65.??:}%, 百拇医药
WELCH, P. J., E. G. MARCUSSON, Q. X. LI, C. BEGER, and M. KRUGER et al., 2000 Identification and validation of a gene involved in anchorage-independent cell growth control using a library of randomized hairpin ribozymes. Genomics 66:274-283.??:}%, 百拇医药
XU, L., C. K. GLASS, and M. G. ROSENFELD, 1999 Coactivator and corepressor complexes in nuclear receptor function. Curr. Opin. Genet. Dev. 9:140-147.??:}%, 百拇医药
ZHENG, P., J. EASTMAN, S. VANDE POL, and S. W. PIMPLIKAR, 1998 PAT1, a microtubule-interacting protein, recognizes the basolateral sorting signal of amyloid precursor protein. Proc. Natl. Acad. Sci. USA 95:14745-14750.(Burt Richards Jon Karpilow Christine Dunn Isaac Peterson Andrew Maxfield Ludmilla Zharkikh Majid Abe)
ABSTRACT(7{-2e., 百拇医药
We used a genetic screening methodology, a human cell line bearing a retinoic-acid-responsive enhanced GFP reporter, and a flow sorter to recover dominant modulators of reporter expression. Four inducers and three suppressors that were fused to the C terminus of a protein scaffold for stability were isolated and their mechanisms of action studied. Mutagenesis experiments indicated that six of these dominant agents exerted their effects at the protein level. The single cDNA coding fragment that was isolated comprised the central 64-amino-acid section of human cyclophilin B, which contained its peptidyl-prolyl isomerase domain; this cyclophilin fragment repressed expression of the retinoic-acid-responsive reporter. The remaining clones encoded peptides shorter than 30 amino acids unrelated to known gene open reading frames. Genetic epistasis studies between the strongest inducer, R3, and a dominant-negative mutant of RAR{alpha} suggest that the two factors function in the same pathway. Transcript microarray analyses suggest that R3 induced a subset of the retinoid-responsive genes in melanoma cells. Finally, yeast two-hybrid assays and co-immunoprecipitation studies of human cell extracts identified PAT1 as a protein that interacts with R3.
DOMINANT genetics has developed over the past several years in an effort to deal with some of the difficulties of screens and selections in cultured mammalian cells, including diploidy and population heterogeneity (HERSKOWITZ 1987 ; KAMB and TENG 2000 ). Using expressed RNA, protein fragments, or peptides (collectively called dominant agents or modulators henceforth), dominant phenotypic modulators have been isolated in several screens, nearly all of which involve selection of survivors from a population of dying or growth-arrested cells (HOLZMAYER et al. 1992 ; GUDKOV et al. 1994 ; AXENOVICH et al. 1998 ; CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; LEVENSON et al. 1999 ; NORMAN et al. 1999 ; XU et al. 1999 ; WELCH et al. 2000 ; BEGER et al. 2001 ; KINSELLA et al. 2002 ). Among the advantages afforded by dominant genetics is its ability to define cellular components, e.g., proteins, which are vunerable to perturbation and the utility of dominant agents as probes to identify their binding targets within cells (GUDKOV et al. 1994 ; CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ).
Here, we use a combination of (i) a fluorescent reporter linked to a promoter; (ii) an optimized clonal cell line engineered with the reporter; and (iii) a flow sorter to isolate, from an expression library, dominant modulators that induce or suppress the expression of the reporter. We investigated the retinoic acid pathway that regulates the expression of genes, many of which contain a retinoic-acid-responsive element (DE LUCA 1991 ; CHAMBON 1996 ). Retinoids regulate gene expression via the RAR-RXR nuclear receptors that act in conjunction with ancillary corepressors and coactivators (reviewed in FREEDMAN 1999 ; XU et al. 1999 ). Retinoids can induce differentiation, growth inhibition, and/or apoptosis in certain cell types and are used to treat specific cancers of the blood and skin diseases (reviewed in SMITH et al. 1992 ; JOHNSON and CHANDRARATNA 1999 ; SAURAT 1999 ; DRAGNEV et al. 2000 ; HANSEN et al. 2000 ; FENAUX et al. 2001 ). Using a retinoic-acid-responsive element-containing promoter to drive a reporter in a human cell line, we identified dominant agents that either induced or suppressed reporter expression.
MATERIALS AND METHODS.mtp, 百拇医药
Generation of expression constructs:.mtp, 百拇医药
A dominant-negative RAR{alpha} receptor, RAR{alpha} {Delta} 403 (TSAI et al. 1992 ), lacking 59 amino acid residues of the C terminus of the wild-type protein, was PCR amplified from a human brain cDNA library. This cDNA fragment was subcloned into a retroviral vector, pVT352, as an in-frame C-terminal fusion to a nonfluorescent variant of green fluorescent protein (dGFP; Tyr-to-Phe mutation at codon 66 in the chromophore). This fusion was placed under the regulation of a constitutive cytomegalovirus promoter; the pVT352 vector carried a neomycin marker for selection of stably transduced cells. One of these scaffolded RAR{alpha} {Delta} 403 clones was verified by DNA sequencing and used in subsequent experiments..mtp, 百拇医药
A random-primed cDNA library of 12 x 106 clones was generated from human placental tissue. Poly(A) mRNA was isolated from placental tissue using standard protocols. The mRNA was used to generate the first strand of cDNA using random nonamer primers linked to a SfiI restriction site (SfiI-A: 5'-GGCCGAGGCGGCC-3'). Second-strand synthesis was accomplished using random hexamer primers linked to a second SfiI restriction site (SfiI-B: 5'-GGCCATAATGGCC-3'). The resulting material was then PCR amplified, size selected, and directionally cloned into compatible SfiI-B to SfiI-A sites of pVT352, where the cDNA fragments were expressed as C-terminal fusions to the nonfluorescent dGFP scaffold.
To make out-of-frame versions of given dominant agents, cDNA fragments were cloned into a variant pVT352 vector that has an altered 3' fusion junction to the dGFP scaffold such that the cDNA sequences present in a different reading frame. All specific expression constructs were subsequently transformed into DH10B Escherichia coli (Life Technologies) and clones were independently sequenced to confirm that desired constructs were generated and that mutations were not introduced. DNAs harvested from sequenced verified clones were then packaged in 293gp cells and the retroviruses produced were used to transduce cells.op!+^gy, 百拇医药
To perform epistasis studies, we replaced the neomycin resistance gene contained in pVT352 with a hygromycin gene from pCMVLacI (Stratagene, La Jolla, CA). The 249-bp R3 sequence was cloned into the SfiI sites downstream of the dGFP and sequence verified. Transduced cells were selected for 8 days in 200 µg/ml hygromycin (Life Technologies).op!+^gy, 百拇医药
Retroviral packaging and transduction:
Infectious retroviral particles were generated by transiently transfecting retroviral plasmid (Moloney murine leukemia virus-based) and VSV-G expression plasmid using LipofectAmine (Life Technologies) into 293gp cells, which stably expressed gag and pol gene products (provided by Inder Verma, Salk Institute). Media from transfected cells was collected after 72 hr and filtered through a 0.45-µm membrane. Transfected cells were grown in Dulbecco's modified Eagle's medium, 10% fetal bovine serum (FBS), and 2 mM L-glutamine (Life Technologies) at 37° in a humidified incubator.ym[89, http://www.100md.com
Transduction of WM35 cells was performed by addition of 20% vol/vol of retroviral supernatant to media and 8 µg/ml polybrene (Sigma, St. Louis) to the media. Cells were exposed to the retrovirus for 24 hr, followed by media change and a 24-hr recovery period prior to selection of retroviral linked selection markers (neomycin, puromycin, or hygromycin). A 20% retroviral supernatant typically resulted in transduction of 30–60% of these human melanoma cells; in some instances, 2% transductions were performed to ensure a single- or low-copy insertion.
WM35 and C8 cells:|{v-)38, http://www.100md.com
Human WM35 melanoma cells were a kind gift from Meenhard Herlyn (Wistar Institute). The WM35 cells were cultured in 80% keratinocyte basal medium and 20% L15 media (Clonetics, San Diego) supplemented with 2% FBS, 2 mM L-glutamine (Life Technologies). The C8 reporter line was generated by fluorescence-activated cell sorter (FACS) and clonal isolations from WM35 cells stably transduced with a retinoic-acid-regulated enhanced GFP reporter (RICHARDS et al. 1999 ). C8 cells were continually grown in media with 2 µg/ml puromycin (Sigma) to maintain the integrated reporter construct.|{v-)38, http://www.100md.com
C8 bioassays:|{v-)38, http://www.100md.com
Two types of bioassays were used to determine the activity of a given dominant agent: bioassay I, which identified modulators that repressed the C8 reporter and thereby made the cells GFP- (dim), and bioassay II, which identified modulators that induced the C8 reporter and thereby made the cells GFP+ (bright).
In bioassay I, C8 cells were transduced with retroviral expression constructs bearing a neomycin marker to introduce desired dominant agents, e.g., RAR{alpha} {Delta} 403. After ">="|%l0, http://www.100md.com
8 days of puromycin and G418 (550 µg/ml) selection, 5 x 105 of these cells were seeded in media with 2% FBS in T75 flasks. Five days later, the cells were trypsinized and analyzed for GFP expression by a Coulter (Hialeah, FL) Epics XL-MCL flow cytometer. Under these conditions, generally 80–95% of the FBS-treated C8 population were GFP+. The GFP- and GFP+ gates were based on the autofluorescence level of WM35 cells, whereby <1% of these naïve cells were present in the GFP+ gate.|%l0, http://www.100md.com
In bioassay II, 4.3 x 105 transduced and stably selected C8 cells were seeded in T75 flasks in media supplemented with 2% charcoal-stripped FBS serum (CBI, Cocalico Biologicals), which minimizes the presence of retinoids. After 5 days, the cells were then trypsinized and analyzed by flow cytometry for GFP expression. The GFP+ gate was set whereby generally <1% of the population from vector-control-transduced C8 cells were present in the gate. For the dose-response studies, C8 cells were maintained in media with 2% CBI for 4 days prior to the addition of given amounts of all-trans retinoic acid (ATRA; Sigma) dissolved in 100% ethanol. The cells were analyzed by flow cytometry 2 days after induction by ATRA.
Viability of cells in all bioassays was determined by propidium iodide staining. All analysis was performed on a single-cell viable subpopulation based on the flow cytometry of forward- and side-scatterplots (RICHARDS et al. 1999 ). In all data shown, the amount of cell death was generally <7%. The GFP signal was detected using an FL1 photomultiplier (525 ± 10 nm); voltages that clearly discriminated between the WM35 autofluorescence signal and the GFP+ signal were set.4lx}d[&, http://www.100md.com
Cycling of libraries:4lx}d[&, http://www.100md.com
Throughout the screens, desired GFP+ or GFP- cells were isolated by using a Coulter EPICS Elite fluorescence-activated cell sorter. The number of C8 cells transduced with retroviral cDNA libraries at each cycle allowed for more than fivefold coverage of the libraries. Genomic DNAs from WM35 cells of interest were isolated using a DNA kit from QIAGEN (Chatsworth, CA). Recombinant cDNA sequences introduced within the retroviral expression vector were recovered by PCR using oligonucleotides 5'-GGATCACTCTCGGCATGGACGAG-3' and 5'-ATTTTATCGAT GTTAGCTTGGCCATT-3', which flanked the cDNA fragment. In general, PCR was performed for 23 cycles under the following conditions: 94° for 30 sec, 68° for 2.5 min. The amplification reactions contained 10 mM KCl, 20 mM Tris-HCL, pH 8.8, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Triton X-100, 0.25 mM dNTPs, 1 µM primers, 1 µg of genomic DNA, and 0.1 units/µl Vent DNA polymerase (New England Biolabs, Beverly, MA). PCR products were digested with SfiI, directionally recloned into the pVT352 retroviral vector, and transformed into DH10B E. coli. DNA was then isolated and packaged in 293gp cells. Naïve C8 cells were transduced with these retroviruses and the neomycin and puromycin selected cells were reexamined in bioassay I or II.
Immunoprecipitation and Western analyses:a/8;}, 百拇医药
Immunoprecipitation experiments were performed by transiently cotransfecting PAT1 and R3 expression constructs into HEK293 cells by LipofectAmine using the manufacturer's protocol (Life Technologies). Two days post-transfection cells were rinsed in PBS, counted, and lysed in immunoprecipitation (IP) buffer: 25 mM HEPES (pH 7.5), 5 mM EDTA, 5 mM EGTA, 50 mM NaCl, 50 mM NaF, 10% (vol/vol) glycerol, 1% (vol/vol) Triton X-100, 2 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (Sigma Aldrich). The lysate was precleared followed by addition of 1 µg of anti-GFP monoclonal antibody (Roche Molecular Biochemicals) and incubation at 4° for 2 hr. Protein G plus Sepharose beads (Amersham Pharmacia Biotech) were subsequently added for 2 hr, followed by a wash with IP buffer. Agarose-protein complexes were resuspended in sample loading buffer processed for silver staining (Invitrogen, San Diego) and PAT1 Western blotting as previously described (ZHENG et al. 1998 ) using anti-PAT1 antibodies generously provided by Dr. Sanjay Pimplikar (Case Western Reserve University).
Immunoblots were performed on WM35 cells infected with each cDNA retroviral supernatant. The parental WM35 cell line was chosen because it does not express any GFP so fusions could be detected unambiguously using anti-GFP antibody. Western blots were carried out using standard protocols (HARLOW and LANE 1988 ). Briefly, total cell lysate from ~ 2 x 105 cells was loaded per well and separated at 200 V on a 10% Tris-Bis gel (Novex, San Diego). The proteins were then transferred to polyvinylidene fluoride membrane for 3 hr at 30 V. A monoclonal GFP antibody (Zymed, S. San Francisco) diluted 1/1000 was used for 1 hr at room temperature followed by a goat anti-mouse HRP-conjugated antibody diluted 1/5000. Proteins were indirectly detected using an enhanced chemiluminescence kit (Amersham).3p, 百拇医药
Gene expression profiling:3p, 百拇医药
Total RNA and poly(A) RNA was isolated from ~ 50 x 106 cells according to the protocols of QIAGEN. Messenger RNAs from two different C8 populations, test (R3 expressing or ATRA treated) and reference (R3of-expressing cells maintained in CBI) were labeled with either CY3 or CY5 and then competitively hybridized to a UniGEM V5 human microarray by Incyte Genomics. Detection and analysis of the ratio of transcript signals observed for the CY3 vs. CY5 probes were performed by Incyte. The calculated balanced differential expression (BDE) for each gene represents a ratio of test and reference signals for the microarray experiment. Genes that were induced (test signal greater than reference signal) were given a positive BDE number (+ test/reference), whereas repressed genes (test signal less than reference signal) were given a negative BDE number (- reference/test).
Yeast two-hybrid assays:^w, 百拇医药
To identify potential R3 interactors, a Lex-dGFP-R3 bait was used in a two-hybrid screen against three commercial libraries: Proquest fetal brain library (Invitrogen), Proquest HeLa library (Invitrogen), and Lex fetal brain library (BD Biosciences Clontech). yVT87 (EGY48) [MAT{alpha} , ura3, his3, trp1, LexAop(X6)-LEU2] yeast cells were transformed with pVT2527 that expresses dGFP-R3 fused to the LexA DNA-binding domain (LexA-dGFP-R3). The resulting strain was mated overnight in rich media (YEPD) to yVT99 [MATa, ura3, his3, trp1, LexAop(X6)-LEU2, LexAop(X8)-URA3] yeast harboring the two-hybrid cDNA libraries. The diploids were first selected on diploid selection media, synthetic-glucose plates lacking histidine and tryptophan, and then screened on interaction selection media, synthetic-glucose plates lacking histidine, tryptophan, uracil, and leucine for the Proquest libraries and on synthetic-galactose plates lacking histidine, tryptophan, uracil, and leucine for the Lex fetal brain library. The library coverage was at least threefold for each screen.
The validity of the yeast two-hybrid interaction was confirmed by a retest whereby plasmid DNA was isolated from colonies that emerged on the selection plates, propagated in E. coli, and retransformed into yVT99 yeast. These cells were then mated to yVT87 yeast harboring LexA-dGFP-R3, LexA-dGFP, or LexA-R3. The diploids were grown on synthetic-glucose plates lacking histidine and tryptophan and replica plated onto diploid selection media and interaction selection media. Increased growth on interaction selection media with LexA-dGFP-R3 compared to LexA-dGFP suggested that the interaction involved the R3 peptide. Several clones (e.g., BIP GenBank accession no. ; SFRS2, NM_003016) displayed interaction with the LexA-dGFP control, indicating that their readout did not require the presence of the R3 peptide; these define a class of false positives that associated with the bait scaffold. Another class of false positives was defined by "sticky" proteins that are commonly isolated from yeast two-hybrid screens, e.g., proteasome subunits and heat-shock proteins (see ).
RESULTS60y@f, 百拇医药
A retinoic-acid-responsive reporter line:60y@f, 百拇医药
Previously, we reported generation of a clonal human cell line, named C8, stably transduced with a retinoid-responsive-enhanced GFP reporter (RICHARDS et al. 1999 ). As shown in , when maintained in media supplemented with CBI (charcoal-stripped FBS to remove retinoids), 1–5% of the C8 population were present in the GFP+ gate; in contrast, 80–95% of these cells induced GFP expression (GFP+) upon treatment with 2% FBS. To further characterize the behavior of this retinoic-acid-regulated reporter, we performed a dose-response study in which C8 cells were treated with concentrations of ATRA ranging from 50 pM to 5 µM . The fluorescence histograms for this series of ATRA doses were essentially bimodal, producing a GFP- (dim) peak and a GFP+ (bright) peak. As the dose of ATRA increased, a greater fraction of induced GFP+ cells was detected while the fluorescence X-means remained relatively similar. The dose-response plot revealed that maximal induction occurred at >500 nM ATRA while the lower limit of reporter activation was detected at ~ 50 pM. Treatment with 5 pM concentrations of ATRA did not induce the reporter (data not shown). The bimodal response of the C8 reporter in human melanoma cells demonstrates that the cis- and trans-regulatory elements involved in this retinoic-acid-responsive system convert a graded ligand input into an all-or-none transcriptional readout on a single cell level. Similar analog-to-digital conversions (or threshold responses) have previously been described for other mammalian promoters (FIERING et al. 1990 ; KARTTUNEN and SHASTRI 1991 ; ROSSI et al. 2000 ).
fig.ommitted^istrw], http://www.100md.com
Figure 1. Isolation of a retinoid/serum-responsive clone. (A) Fluorescence histograms of C8 cell populations showing the number of cells on the y-axis and relative fluorescence units (RFU) on the x-axis. The C8 populations transduced with control pVT352 retroviral plasmid were maintained for 5 days in media supplemented with FBS (green) or CBI (red). Also shown are C8 cells transduced with the dominant-negative RAR403 construct that were subsequently maintained in FBS for 5 days before analysis by flow cytometry (blue). (B) A phase-contrast field of RAR403 transduced C8 cells maintained in media with FBS. (C) The same field demonstrating a speckled phenotype of GFP+ and GFP- cells as observed by fluorescence microscopy. The arrowheads indicate some cells that do not express GFP.^istrw], http://www.100md.com
fig.ommitted^istrw], http://www.100md.com
Figure 2. All-or-none response of the C8 reporter. Shown are the fluorescence histograms of C8 cells. The cells were initially maintained in media with CBI for 3 days, followed by treatment with the indicated concentrations of ATRA and flow cytometry analyses 2 days later. In the CBI-only control, no ATRA was added. Induction of the retinoic-acid-regulated GFP construct in these cells produces a bimodal distribution consisting of a GFP- population and a GFP+ population. Concentrations of ATRA >5 µM did not increase the fraction of cells in the GFP+ gate beyond that observed for the 5 µM ATRA sample (data not shown). The higher percentage of GFP+ cells in the FBS vs. the ">="
5 µM ATRA populations suggests that other factors besides retinoids can contribute to the activation of the C8 reporter. For comparison, the autofluorescence of parental WM35 melanoma cells lacking the reporter construct is presented in the top box. The GFP- gate was based on the autofluorescence level of naïve WM35 cells in which <1% of the population was present in the GFP+ gate.54nrul, 百拇医药
Modulation by a dominant-negative RAR{alpha} mutant:54nrul, 百拇医药
To validate the C8 reporter, we expressed RAR403, a known dominant-negative inhibitor of the RAR-RXR transcription complex, in C8 cells (TSAI et al. 1992 ; ISOGAI et al. 1997 ). We fused RAR403 cDNA to a nonfluorescent GFP scaffold (dGFP), chosen for its ability to enhance stability of fused peptides (ABEDI et al. 1998 ; SANDROCK et al. 2001 ), and expressed the fusion using a constitutive cytomegalovirus promoter in a retroviral vector. In contrast to parental C8 cells, C8 derivatives stably transduced with RAR403 retrovirus exhibited a bimodal distribution with 35% of the cells in the GFP- gate after 5 days of growth in media with FBS . This bimodal response likely reflected the penetrance of RAR403 in C8 cells because individual clones of RAR403-expressing cells had the same bimodal behavior (data not shown). When examined by fluorescence microscopy, single C8:RAR403 colonies displayed a "speckled" phenotype due to the mixture of GFP+ and GFP- cells in the colony . The repressor effect of RAR403 was consistent with its proposed mutant activity; i.e., it dimerizes with native RXRs and associates with retinoic-acid-responsive elements and corepressors to shut down transcription in a dominant fashion (TSAI et al. 1992 ). These data demonstrated that the C8 reporter responds as expected to known modulators of the RAR-RXR transcription machinery.
Repressors of C8 reporter expression:%km-jo%, 百拇医药
Using C8 cells, we searched for dominant agents that modulated expression of the retinoic-acid-responsive reporter. Two screens were performed: one to identify agents that repressed the C8 reporter and made the cells GFP-, or dim (class I), and one to identify agents that induced the C8 reporter and made the cells GFP+, or bright (class II).%km-jo%, 百拇医药
To find class I modulators, we expressed a human placental cDNA library fused to a nonfluorescent dGFP scaffold in C8 cells. The transduced C8 cells were maintained in media with FBS for 5 days (bioassay I) and the dimmest 5% of the F0 population was collected by FACS . The F1 subpopulation was expanded and subjected to two further cycles of dim sorting after which ~ 75% of cells were present in the GFP- gate , F3 subpopulation). Because colonies of C8 cells transduced with RAR403 exhibited a speckled appearance, we screened F3 cells for a similar phenotype. Colonies originating from single F3 cells were maintained in FBS for 5 days. When examined by fluorescence microscopy, ~ 10% of the F3 colonies were essentially GFP+ while ~ 40% of the colonies were GFP-. The fully bright colonies were presumably derived from cells with an intact retinoid/serum-response pathway. On the basis of control experiments, most of the completely dim colonies likely originated from cells with inactivating mutations in the retinoic-acid-regulated GFP retroviral reporter or with lesions at loci required to transactivate the reporter (data not shown).
fig.ommitted11x&, 百拇医药
Figure 3. Class I modulators that repress the C8 reporter. (A) Histogram profiles of C8 cells grown in 2% FBS for library sorts one through three. Approximately 25 x 106 C8 cells were transduced with the placental cDNA library retrovirus and selected for 8 days in G418-containing media. To isolate dominant agents that repressed the reporter, the cells were transferred into media with 2% FBS and, after 5 days, cells in the GFP- gate were recovered by FACS. These F1 GFP- sorted C8 cells were expanded and similarly sorted two more times to generate the F2 and F3 subpopulations. (B) C8 cells were transduced with the following constructs: vector, RAR403, and the in-frame (solid) and out-of-frame (of; stippled) versions of F791, F802, and F820; these were examined in bioassay I. (C) Flow cytometry data for vector (control) and the F820 in-frame and out-of-frame (F820of) constructs shown as representative data for the class I modulators. (D) Western analysis of WM35 cells transduced with each construct to confirm protein expression. With the exception of the RAR403 clone, each of the constructs expressed approximately similar protein levels and exhibited the expected size band based on the number of amino acid residues fused to the GFP scaffold (see ). The RAR403 clone was expressed at the mRNA level as detected by RT-PCR (data not shown). A nonspecific 60-kD band, recognized in all samples, serves as a control for equivalent loading.
Complementary DNAs from the F3 sublibrary were recovered from individual speckled colonies and a group of 10 randomly selected inserts were analyzed further. Relative to the vector control, 3 of the 10 clones (F797, F802, and F820) caused >36% of cells to fall within the GFP- gate . For comparison, 41% of dominant-negative RAR403-transduced C8 cells were in the GFP- gate. The bimodal fluorescence histogram of C8 cells containing F820 is shown as representative data .s'h-hun, 百拇医药
To determine if the three class I modulators acted at the protein or RNA level, each of the three sequences was expressed in an alternate reading frame (out of frame, of;. In the case of clone F797, the in-frame and out-of-frame constructs displayed a similar increase in cells in the GFP- gate, suggesting RNA-level activity. In contrast, the out-of-frame constructs of F802 and F820 did not produce a significant elevation in the percentage of GFP- cells above the vector control. To confirm that these out-of-frame constructs produce protein (and by inference, mRNA), we performed Western blots with an anti-GFP antibody. Each of the constructs expressed approximately equivalent amounts of protein at the expected size. Taken together, these results suggested that F802 and F820 act at the protein level. Of interest, expression of RAR403 was not detectable on this immunoblot even though C8 reporter repression was observed in cells transduced with this construct . This result could be explained by the observation that a high level of expression of the dominant-negative RAR{alpha} polypeptide is cytotoxic in these human cells (data not shown).
The identities of the dominant agents were investigated by sequence comparisons with public nucleotide databases. Matches were identified for F797 and F802, but not F820 (. The F797 clone consisted of 572 nucleotides from the sense strand of the HSP90 mRNA 5' untranslated region. Clone F802 encoded 64 amino acid residues from the central portion of cyclophilin B, including the peptidyl-prolyl cis-trans isomerase domain, fused in-frame with the dGFP scaffold. The F820 sequence consisted of a 249-base insert that encoded a 29-amino-acid peptide fused to dGFP.w(:-0k, 百拇医药
fig.ommittedw(:-0k, 百拇医药
Table 1. Dominant modulators of retinoid signalingw(:-0k, 百拇医药
Inducers of C8 reporter expression:w(:-0k, 百拇医药
In addition to identifying class I modulators, we screened for class II modulators that induced the C8 reporter and thereby produced GFP+ cells. C8 cells harboring the expression library were maintained in media with CBI for 5 days (bioassay II), after which the brightest 0.2% of the population was collected by FACS. This subpopulation was grown and genomic DNA was isolated from the cells. The retroviral inserts were recovered by PCR and amplified fragments were recloned into retroviral expression vector, packaged, and used to transduce naïve C8 cells. After neomycin selection, the GFP+ subpopulation was again recovered by FACS from cells maintained in CBI. This cycling process was repeated five more times. By the sixth cycle, the fraction of cells in the GFP+ gate was 23-fold higher than that of control cells .
fig.ommitted[qk?'x, 百拇医药
Figure 4. Class II modulators that induce the C8 reporter. (A) A comparison of the percentage of GFP+ cells for library- and vector-control-transduced C8 cells. In the first step, ~ 30 x 106 C8 cells were transduced with the placental cDNA library retrovirus and selected for 8 days in G418-containing media. To isolate dominant agents that induced the retinoic-acid-regulated GFP reporter, the cells were transferred into media with 2% CBI and, after 5 days, cells in the GFP+ gate were recovered by FACS. Retroviral cDNA inserts from the sorted cells were recovered by PCR, recloned into retroviral vector, packaged, and used to transduce naïve C8 cells. The GFP+ cells in this sublibrary-transduced C8 population were isolated by FACS. (B) A histogram of cycle-6-transduced C8 cells compared to vector-control-transduced cells. (C) C8 cells were transduced with the following constructs: control vector and the in-frame (solid) and out-of-frame (of; stippled) versions of R1, R2, R3, and R44, which were examined in bioassay II. (D) Flow cytometry data for vector and R3 in-frame and out-of-frame (R3of) constructs shown as representative data for the class II modulators. (E) Western analysis of WM35 cells transduced with each construct to confirm protein expression. The R1of clone was expressed at the mRNA level as detected by reverse-transcription-based PCR (data not shown).
We sequenced 134 clones from the cycle 6 sublibrary to ascertain identities and representation. The most abundant clone, R3, comprised 37% of sequences analyzed, while the next most abundant clones, R1 and R2, accounted for 13 and 9% of the sequences, respectively. We tested the 15 most frequent clones independently in C8 bioassay II and found that 4 of them exhibited significant signals above the vector control. The strongest effect was elicited by the R3 clone: 54% of the R3-infected C8 cells were present in the GFP+ gate compared to 1% in control cells . Sequence analysis revealed that the R1, R2, R3, and R44 clones encoded peptides of nine residues or fewer fused to the dGFP scaffold . The possibility that the peptide fusions of these clones to the dGFP scaffold made the mutant protein fluorescent was eliminated by showing that expression of these clones in WM35 or other cells did not produce any detectable fluorescence signal by flow cytometry (data not shown). By comparing the effects of in-frame and out-of-frame constructs and by performing Western analyses for expression , all of the class II modulators were shown to act as peptides. The out-of-frame R1 construct, although it produced significantly less protein compared to R1, expressed similar levels of mRNA as determined by reverse-transcription-based PCR (data not shown).
Mechanism of action of the R3 inducer:ziqy, 百拇医药
We performed genetic epistasis studies to examine the relationship between R3, the strongest class II modulator, and RAR403, a known dominant-negative factor of the retinoid pathway. Using bioassay I as the readout, we found that a lower percentage of GFP- cells was present in the population expressing both RAR403 and R3 compared to cells expressing RAR403 alone (11 vs. 28%, ). Thus, R3 interfered with the ability of RAR403 to block pathway induction. Using bioassay II as the readout, fewer bright cells were present in the population expressing both proteins than in cells expressing only R3 (22 vs. 35%). Thus, RAR403 partially blocked pathway induction by R3. Taken together, these data indicated that each factor (RAR403 and R3) antagonizes the activity of the other.ziqy, 百拇医药
fig.ommittedziqy, 百拇医药
Figure 5. Genetic epistasis between R3 and dominant-negative RAR{alpha} . C8 cells were initially transduced with either dGFP-R3 fusion or control dGFP (Vector) expressed from a constitutive cytomegalovirus promoter in a retroviral vector carrying a hygromycin selection marker. After selection with hygromycin, some of the cells were secondarily transduced with a retrovirus bearing dGFP-RAR403 and subsequently selected with G418 to generate a population carrying both constructs. These constructs were expressed from a constitutive cytomegalovirus promoter. Shown on the left (solid bars) are the percentages of the GFP- population of C8 cells, transduced with the indicated constructs and subjected to bioassay I (i.e., seeded, maintained in media with 2% FBS, and then analyzed by flow cytometry after 5 days). Shown on the right (hatched) are the percentages of the GFP+ population of C8 cells subjected to bioassay II (i.e., seeded, maintained in media supplemented with 2% CBI, and then analyzed after 6 days). Averaged data for three replicate samples, along with standard deviations, are shown for each construct.
A second approach, gene expression profiling by microarray analyses, was used to compare the effects of R3 on C8 cells to that elicited by ATRA. The primary purpose of these studies was to determine if R3 regulated the expression of a similar set of genes as retinoids. Transcript levels for ~ 9000 genes were compared in C8 cells expressing dGFP-R3 or C8 cells expressing dGFP-R3of that were treated with ATRA. In the case of the R3-expressing cells, there were only 30 genes (outliers) whose expression was up- or downregulated >1.7-fold (; ). These data suggest that R3 causes a limited change in the global expression profile of these cells. In contrast, ATRA treatment of C8 cells expressing R3of caused a more dramatic shift in transcript profile, with 285 genes altered in expression level >1.7-fold (). Notably, the set of 30 genes (>1.7-fold modulated) in R3-expressing cells was essentially a subset of the ATRA-treated data. Overall, 29 of the 30 outliers in the R3 data set responded similarly in the ATRA experiment (+ for induced, - for repressed; ). Moreover, 17 of the 30 outliers in the R3 data set were also outliers of the same sign in the ATRA experiment Covariance calculations revealed that this correlation between the R3 subgroup and the equivalent genes in ATRA-treated cells was highly significant (covariance = 1.89, P << 0.001; ).
fig.ommitted@, http://www.100md.com
Figure 6. Transcript microarray analysis of R3. Scatter plot showing normalized expression levels of 7529 genes (genes not expressed in both samples have been removed) in the following types of C8 cells. (A) Cells expressing dGFP-R3 or dGFP-R3of maintained in CBI. (B) dGFP-R3of cells treated with CBI + ATRA or maintained in CBI alone. Red dots correspond to the subset of outliers of R3 vs. R3of (for which the R3/R3of balanced differential expression is > +1.7 or < -1.7) that are also outliers, with the same sign, in the CBI + ATRA vs. CBI data. (C) Histogram of covariances calculated using the 30-member outlier subgroup for dGFP-R3-expressing cells compared to the same group of genes in ATRA-treated cells (arrow) or to sets of 30 randomly selected genes in the ATRA data set (distribution). Covariance was calculated according to the formula cov(x, y) = (1/N) (xi -
fig.ommittedn9xmar, 百拇医药
Table 2. Outlier genes subgroup of R3 vs. R3of microarray experimentn9xmar, 百拇医药
R3 interactors:n9xmar, 百拇医药
To identify proteins that associate with R3, we executed a yeast two-hybrid screen. dGFP-R3 was fused downstream of a LexA DNA-binding domain and used as a bait to screen activation domain-cDNA libraries derived from fetal human brain tissue and cervical HeLa cells. Briefly, 400 putative positives were isolated from the initial screens. Sequence analyses revealed that 195 of these clones were out of frame or encoded known yeast two-hybrid false positives (see MATERIALS AND METHODS). A total of 205 clones were retested with LexA-dGFP-R3 or LexA-dGFP to verify specificity of the interaction with R3. Nineteen of the clones passed the retest by displaying stronger interaction with the LexA-dGFP-R3 bait compared to the LexA-dGFP control. These 19 positives consisted of partial or full-length in-frame cDNAs encoding four proteins: PAT1 (GenBank accession no. ), HIRIP4 (XM_040816), TCTEL-1 (), and {alpha} -B-crystallin (NM_001885). Specifically, two of the four cDNA clones, TCTEL-1 and {alpha} -B-crystallin, associated weakly with LexA-GFP; importantly, PAT1 and HIRIP did not. In addition, we executed a second study whereby we generated a LexA-R3 bait and used it to ascertain whether any of the four cDNAs interacted with the R3 sequences in a scaffold (GFP)-independent manner. Of the four cDNA clones, only PAT1 displayed interaction with LexA-R3 in the assay , indicating that the nine amino acid residues encoded by R3 were sufficient for enabling interaction with the protein.
fig.ommitted44!, 百拇医药
Figure 7. Interaction of R3 with PAT1. (A) Yeast two-hybrid auxotrophy readout of the indicated baits on a LexA DNA-binding domain with a prey of PAT1 fused to the C terminus of an activation domain (AD). To demonstrate that dGFP-R3 binds to PAT1 in human cells, the two proteins were exogenously co-expressed in HEK293 cells and the resulting soluble fraction of cell extracts was subjected to immunoprecipitation and protein analyses. Data for two independent PAT1 clones, no. 5 and no. 10, are shown. (B) The out-of-frame version of R3, dGFP-R3of, was also analyzed as a control for the specificity of the interaction between GFP-R3 and PAT1. Samples: lane 1, dGFP-R3 and PAT1 clone no. 5; lane 2, dGFP-R3of and PAT1 clone no. 5; lane 3, dGFP-R3 and PAT1 clone no. 10; and lane 4, dGFP-R3of and PAT1 clone no. 10. Specifically, IP was performed using anti-GFP antibodies and the precipitate was examined by (B) silver-stain detection of proteins fractionated by polyacrylamide gel electrophoresis (PAGE) or (C) by Western blotting using anti-PAT1 antibodies. (D) A control Western blot was performed to show equivalent PAT1 expression in the different samples.
Examination of the transcript microarray data from C8 cells revealed that PAT1 was expressed in these human melanoma cells. Specifically, the microarray analyses quantitated signals of 601 arbitrary expression units, 716 units and 595 units in C8 cells treated with ATRA, R3, or R3of, respectively, where <250 units was considered no expression. The microarray data did not show any significant induction or repression (> or <1.7-fold) of PAT1 in the retinoic-acid-induced or R3-expressing C8 cells relative to the R3of-expressing cells.'v[k, 百拇医药
We investigated whether dGFP-R3 and PAT1 interacted in mammalian cells. We initially examined C8 and WM35 cells but found that the expression levels of endogenous PAT1 protein in these melanoma cells were not detectable by Western analyses using anti-PAT1 antibodies; moreover, the cells were not amenable to transfections. Full-length PAT1 was therefore transfected into human HEK293 cells along with either dGFP-R3 or dGFP-R3of. Subsequently, cell extracts were prepared and subjected to IP-Western analyses using anti-GFP antibodies for IP and anti-PAT1 antibodies for the Western blots. As shown in , association between PAT1 and dGFP-R3 was observed in two independent experiments (lanes 1 and 3). In contrast, the out-of-frame R3 construct (lanes 2 and 4) was successfully immunoprecipitated by anti-GFP antibodies but failed to pull down PAT1 . As a control, we demonstrated that PAT1 was equivalently expressed in the different cell extracts
DISCUSSION9v, 百拇医药
Several protein scaffolds, including GFP, thioredoxin, and inactivated staphylococcal nuclease, have been previously used in screens for peptide aptamers that evoke dominant effects in cellular processes (COLAS et al. 1996 ; CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ). The primary reason for using a protein scaffold for the presentation of peptides is stability, as studies have shown that high levels of expression of a scaffolded peptide can be achieved within cells (SANDROCK et al. 2001 ). In addition, the scaffold serves as a tag for the dominant agent, e.g., for monitoring expression, subcellular localization, or purification. However, the use of a scaffold in this type of context raises other issues. For example, the scaffold is sizable (30 kD in the case of GFP) and may thus stearically hinder the association of peptide fusions to certain proteins. Although the scaffold is assumed to be "inert," it may not be so in certain instances. It is conceivable that the scaffold contributes to the association of the peptide fusion with its target. Moreover, once bound to a target protein, the scaffold could occlude the binding of other molecules that normally interact with the protein.
All of the dominant agents from our screen were fused to the C terminus of the dGFP scaffold. By comparing the activity and expression levels of in- and out-of-frame versions of the clones, we demonstrate that the activities of six of the seven dominant agents reported here reside in the expressed protein. Only the F797 clone appeared to act at the RNA level. Of the isolated clones, F802 was the sole one that encoded a fragment of a known protein (cyclophilin B). The other five peptide modulators consisted of peptides ranging from 3 to 29 amino acid residues. The low frequency of protein fragments obtained from the screen is primarily due to the fact that the random-primed cDNA expression library used in these selections is expected to contain a majority of clones that express peptides derived from dGFP fusions to out-of-frame coding or untranslated sequences.+./3n, 百拇医药
We chose to investigate the R3 peptide most thoroughly because of the strength of its effect, and because it mimics ATRA action and therefore has potential relevance to retinoid-based therapies such as promyelocytic leukemia treatment (SMITH et al. 1992 ; VARGHESE et al. 1999 ; FENAUX et al. 2001 ). Through genetic epistasis studies with a dominant-negative RAR{alpha} mutant , we demonstrated that R3 acts within the retinoid pathway. Analysis of the microarray data suggests that R3 activity is a subset of the biological effects of ATRA (). R3 appears to have striking specificity; only a single gene from the 30-member outlier group in R3 vs. R3of was negatively correlated with the corresponding gene in ATRA-treated cells (). Thus, the phenotypic consequences of R3 expression appear to be largely those that were selected using the retinoic-acid-responsive reporter, namely, induction of the pathway.
Many other dominant peptide aptamers that have been characterized in other systems have been shown to function as inhibitors of protein function (COLAS et al. 1996 ; CAPONIGRO et al. 1998 ; KOLONIN and FINLEY 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ; BLUM et al. 2000 ; BUTZ et al. 2000 ). It is therefore plausible that R3 inhibits the activity of a repressor to induce reporter expression in C8 cells. Other mechanistic explanations for R3 action were not supported by the data. For example, expression of R3 did not alter fluorescence in WM35 cells that express enhanced GFP from a constitutive SV40 promoter (data not shown). This result excludes the possibility that R3 stabilizes the GFP mRNA or protein.1t6%[!, 百拇医药
As previously shown, if a peptide acts on another protein to elicit its effect, protein-protein interaction assays such as the yeast two-hybrid system can be used to identify candidate targets (CAPONIGRO et al. 1998 ; GEYER et al. 1999 ; NORMAN et al. 1999 ). Using yeast two-hybrid assays, we showed that coding segments from PAT1, HIRIP4, TCTEL-1, and {alpha} -B-crystallin interacted with a Lex-dGFP-R3 bait. PAT1 and HIRIP4 did not interact with the Lex-dGFP control bait, indicating that the nine C-terminal amino acid residues encoded by R3 were necessary for interaction with these cDNA clones. Moreover, only PAT1 interacted with R3 in a dGFP-independent manner in the yeast two-hybrid assay and co-immunoprecipitation studies showed that PAT1 binds to dGFP-R3 in human cells . PAT1 was initially described by ZHENG et al. 1998 as a kinesin-like protein that associates with the basolateral sorting signal of amyloid precursor protein, as well as with microtubules and membranes. The primary sequence of PAT1 contains numerous motifs of interest, including four kinesin light-chain-like repeats, two leucine zipper domains, a "LXLL" motif (NR box) found in transcriptional coregulators, a PEST protein-degradation sequence, as well as several putative nuclear localization signals. GAO and PIMPLIKAR 2001 have reported that PAT1 degradation is augmented by the expression of certain APP proteolytic fragments, which also represses retinoic-acid-responsive gene expression. Intriguingly, this observation is conversely consistent with preliminary studies that reveal that GFP-R3, an inducer of the retinoid pathway, increases the levels of PAT1 protein in mammalian cells (S. PIMPLIKAR, personal communication). Further work is therefore warranted to elucidate any role that PAT1 may have in retinoid signaling.
It is less certain how the other dominant agents act. The F797 RNA-level modulator cannot function as a classic antisense molecule because it matches the sense strand of HSP90. It could behave as a cross-hybridizing antisense RNA or as an RNA aptamer (FIRE 1999 ; HERMANN and PATEL 2000 ). The three amino acid peptide modulators (R1, R2, and R44) are of special interest because of their size. These dominant agents may act at least partly through some residues of the dGFP scaffold. F802 consists of the central peptidyl-prolyl isomerase domain of cyclophilin B (CypB), a member of a family of enzymes that influences protein structure and function (PRICE et al. 1991 ). Native CypB mediates several cellular processes (SWANSON et al. 1992 ; RYCYZYN et al. 2000 ), possesses an N-terminal signal sequence, and has been reported to be secreted (ALLAIN et al. 1999 ). Since F802 appears to be localized within the cell, it seems unlikely that F802 functions in a classic dominant-negative manner, unless CypB has intracellular activities as well. It is conceivable that F802 interacts with one or more of the ligands that another cyclophilin binds and thereby interferes with its function.
CONCLUSIONS#ldm, 百拇医药
Using the optimized retinoic-acid-responsive C8 reporter line, we executed two en masse genetic screens that identified repressors and inducers of the reporter. Six of seven of these modulators functioned as peptides. A single clone comprised the coding fragment of a known gene, CypB. Initial transcript microarray and genetic epistasis studies on the mechanism of action of R3 strongly supported its role as a specific inducer of the retinoid pathway. PAT1 was identified as a protein that interacts with R3. The dominant agents isolated in this study should be useful as reagents to probe and further understand the underpinnings of retinoid signaling. Moreover, because retinoids are used for the treatment of certain leukemias and skin diseases, and because the R3 peptide appears to induce the pathway in a manner similar to ATRA, it will be of interest to determine if R3 or its target has therapeutic utility.#ldm, 百拇医药
ACKNOWLEDGMENTS
We are indebted to Dr. Nikunj Somia for sharing his expertise on retroviruses and Dr. Sanjay Pimplikar for information about PAT1 and for kindly providing antibodies against the protein. We are grateful to Anthony Gaglio, Sharon Malmstrom, Stephen Hansen, and Rex Malmstrom for excellent technical assistance. We wish to thank Drs. Giordi Caponigro, Michael Feldhaus, Michael Pierce, Mark Poritz, and Mahendra Rao for input throughout this project and for comments on the manuscript.tx, 百拇医药
Manuscript received October 4, 2002; Accepted for publication December 3, 2002.tx, 百拇医药
LITERATURE CITEDtx, 百拇医药
ABEDI, M. R., G. CAPONIGRO, and A. KAMB, 1998 Green fluorescent protein as a scaffold for intracellular presentation of peptides. Nucleic Acids Res. 26:623-630.tx, 百拇医药
ALLAIN, F., S. DURIEUX, A. DENYS, M. CARPENTIER, and G. SPIK, 1999 Cyclophilin B binding to platelets supports calcium-dependent adhesion to collagen. Blood 94:976-983.tx, 百拇医药
AXENOVICH, S. A., A. R. KAZAROV, A. D. BOIKO, G. ARMIN, and I. B. RONINSON et al., 1998 Altered expression of ubiquitous kinesin heavy chain results in resistance to etoposide and hypersensitivity to colchicine: mapping of the domain associated with drug response. Cancer Res. 58:3423-3428.
BEGER, C., L. N. PIERCE, M. KRUGER, E. G. MARCUSSON, and J. M. ROBBINS et al., 2001 Identification of Id4 as a regulator of BRCA1 expression by using a ribozyme-library-based inverse genomics approach. Proc. Natl. Acad. Sci. USA 98:130-135.t4-^s;^, 百拇医药
BLUM, J. H., S. L. DOVE, A. HOCHSCHILD, and J. J. MEKALANOS, 2000 Isolation of peptide aptamers that inhibit intracellular processes. Proc. Natl. Acad. Sci. USA 97:2241-2246.t4-^s;^, 百拇医药
BUTZ, K., C. DENK, A. ULLMANN, M. SCHEFFNER, and F. HOPPE-SEYLER, 2000 Induction of apoptosis in human papillomaviruspositive cancer cells by peptide aptamers targeting the viral E6 oncoprotein. Proc. Natl. Acad. Sci. USA 97:6693-6697.t4-^s;^, 百拇医药
CAPONIGRO, G., M. R. ABEDI, A. P. HURLBURT, A. MAXFIELD, and W. JUDD et al., 1998 Transdominant genetic analysis of a growth control pathway. Proc. Natl. Acad. Sci. USA 95:7508-7513.t4-^s;^, 百拇医药
CHAMBON, P., 1996 A decade of molecular biology of retinoic acid receptors. FASEB J. 10:940-954.t4-^s;^, 百拇医药
COLAS, P., B. COHEN, T. JESSEN, I. GRISHINA, and J. MCCOY et al., 1996 Genetic selection of peptide aptamers that recognize and inhibit cyclin-dependent kinase 2. Nature 380:548-550.
DE LUCA, L. M., 1991 Retinoids and their receptors in differentiation, embryogenesis, and neoplasia. FASEB J. 5:2924-2933.##n7$4j, 百拇医药
DRAGNEV, K. H., J. R. RIGAS, and E. DMITROVSKY, 2000 The retinoids and cancer prevention mechanisms. Oncologist 5:361-368.##n7$4j, 百拇医药
FENAUX, P., C. CHOMIENNE, and L. DEGOS, 2001 All-trans retinoic acid and chemotherapy in the treatment of acute promyelocytic leukemia. Semin. Hematol. 38:13-25.##n7$4j, 百拇医药
FIERING, S., J. P. NORTHROP, G. P. NOLAN, P. S. MATTILA, and G. R. CRABTREE et al., 1990 Single cell assay of a transcription factor reveals a threshold in transcription activated by signals emanating from the T-cell antigen receptor. Genes Dev. 4:1823-1834.##n7$4j, 百拇医药
FIRE, A., 1999 RNA-triggered gene silencing. Trends Genet. 15:358-363.##n7$4j, 百拇医药
FREEDMAN, L. P., 1999 Increasing the complexity of coactivation in nuclear receptor signaling. Cell 97:5-8.##n7$4j, 百拇医药
GAO, Y. and S. W. PIMPLIKAR, 2001 The gamma-secretase-cleaved C-terminal fragment of amyloid precursor protein mediates signaling to the nucleus. Proc. Natl. Acad. Sci. USA 98:14979-14984.
GEYER, C. R., A. COLMAN-LERNER, and R. BRENT, 1999 "Mutagenesis" by peptide aptamers identifies genetic network members and pathway connections. Proc. Natl. Acad. Sci. USA 96:8567-8572.in;$, 百拇医药
GUDKOV, A. V., A. R. KAZAROV, R. THIMMAPAYA, S. A. AXENOVICH, and I. A. MAZO et al., 1994 Cloning mammalian genes by expression selection of genetic suppressor elements: association of kinesin with drug resistance and cell immortalization. Proc. Natl. Acad. Sci. USA 91:3744-3748.in;$, 百拇医药
HANSEN, L. A., C. C. SIGMAN, F. ANDREOLA, S. A. ROSS, and G. J. KELLOFF et al., 2000 Retinoids in chemoprevention and differentiation therapy. Carcinogenesis 21:1271-1279.in;$, 百拇医药
HARLOW, E., and D. LANE, 1988 Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.in;$, 百拇医药
HERMANN, T. and D. J. PATEL, 2000 Adaptive recognition by nucleic acid aptamers. Science 287:820-825.in;$, 百拇医药
HERSKOWITZ, I., 1987 Functional inactivation of genes by dominant negative mutations. Nature 329:219-222.
HOLZMAYER, T. A., D. G. PESTOV, and I. B. RONINSON, 1992 Isolation of dominant negative mutants and inhibitory antisense RNA sequences by expression selection of random DNA fragments. Nucleic Acids Res. 20:711-717.-d#6e8, 百拇医药
ISOGAI, M., M. V. CHIANTORE, M. HAQUE, G. SCITA, and L. M. DE LUCA, 1997 Expression of a dominant-negative retinoic acid receptor construct reduces retinoic acid metabolism and retinoic acid-induced inhibition of NIH-3T3 cell growth. Cancer Res. 57:4460-4464.-d#6e8, 百拇医药
JOHNSON, A. and R. A. CHANDRARATNA, 1999 Novel retinoids with receptor selectivity and functional selectivity. Br. J. Dermatol. 140(Suppl. 54):12-17.-d#6e8, 百拇医药
KAMB, A. and D. H. TENG, 2000 Transdominant genetics, peptide inhibitors and drug targets. Curr. Opin. Mol. Ther. 2:662-669.-d#6e8, 百拇医药
KARTTUNEN, J. and N. SHASTRI, 1991 Measurement of ligand-induced activation in single viable T cells using the lacZ reporter gene. Proc. Natl. Acad. Sci. USA 88:3972-3976.-d#6e8, 百拇医药
KINSELLA, T. M., C. T. OHASHI, A. G. HARDER, G. C. YAM, and W. LI et al., 2002 Retrovirally delivered random cyclic peptide libraries yield inhibitors of interleukin-4 signaling in human B cells. J. Biol. Chem. 277:37512-37518.
KOLONIN, M. G. and R. L. FINLEY, JR., 1998 Targeting cyclin-dependent kinases in Drosophila with peptide aptamers. Proc. Natl. Acad. Sci. USA 95:14266-14271.79kwq], http://www.100md.com
LEVENSON, V. V., E. LAUSCH, D. J. KIRSCHLING, E. V. BROUDE, and I. A. DAVIDOVICH et al., 1999 A combination of genetic suppressor elements produces resistance to drugs inhibiting DNA replication. Somat. Cell Mol. Genet. 25:9-26.79kwq], http://www.100md.com
NORMAN, T. C., D. L. SMITH, P. K. SORGER, B. L. DREES, and S. M. O'ROURKE et al., 1999 Genetic selection of peptide inhibitors of biological pathways. Science 285:591-595.79kwq], http://www.100md.com
PRICE, E. R., L. D. ZYDOWSKY, M. J. JIN, C. H. BAKER, and F. D. MCKEON et al., 1991 Human cyclophilin B: a second cyclophilin gene encodes a peptidyl-prolyl isomerase with a signal sequence. Proc. Natl. Acad. Sci. USA 88:1903-1907.79kwq], http://www.100md.com
RICHARDS, B., J. KARPILOW, C. DUNN, L. ZHARKIKH, and A. MAXFIELD et al., 1999 Creation of a stable human reporter cell line suitable for FACS-based, transdominant genetic selection. Somat. Cell Mol. Genet. 25:191-205.
ROSSI, F. M., A. M. KRINGSTEIN, A. SPICHER, O. M. GUICHERIT, and H. M. BLAU, 2000 Transcriptional control: rheostat converted to on/off switch. Mol. Cell 6:723-728.p, http://www.100md.com
RYCYZYN, M. A., S. C. REILLY, K. O'MALLEY, and C. V. CLEVENGER, 2000 Role of cyclophilin B in prolactin signal transduction and nuclear retrotranslocation. Mol. Endocrinol. 14:1175-1186.p, http://www.100md.com
SANDROCK, T. M., B. RISLEY, B. T. RICHARDS, M. A. PORITZ, and H. A. AUSTIN et al., 2001 Exogenous peptide and protein expression levels using retroviral vectors in human cells. Mol. Ther. 4:398-406.p, http://www.100md.com
SAURAT, J. H., 1999 Retinoids and psoriasis: novel issues in retinoid pharmacology and implications for psoriasis treatment. J. Am. Acad. Dermatol. 41:S2-S6.p, http://www.100md.com
SMITH, M. A., D. R. PARKINSON, B. D. CHESON, and M. A. FRIEDMAN, 1992 Retinoids in cancer therapy. J. Clin. Oncol. 10:839-864.p, http://www.100md.com
SWANSON, S. K., T. BORN, L. D. ZYDOWSKY, H. CHO, and H. Y. CHANG et al., 1992 Cyclosporin-mediated inhibition of bovine calcineurin by cyclophilins A and B. Proc. Natl. Acad. Sci. USA 89:3741-3745.
TSAI, S., S. BARTELMEZ, R. HEYMAN, K. DAMM, and R. EVANS et al., 1992 A mutated retinoic acid receptor-alpha exhibiting dominant-negative activity alters the lineage development of a multipotent hematopoietic cell line. Genes Dev. 6:2258-2269.??:}%, 百拇医药
VARGHESE, L., A. JANCKILA, and L. T. YAM, 1999 Acute promyelocytic leukemia: new methods in diagnosis and treatment. J. Ky. Med. Assoc. 97:61-65.??:}%, 百拇医药
WELCH, P. J., E. G. MARCUSSON, Q. X. LI, C. BEGER, and M. KRUGER et al., 2000 Identification and validation of a gene involved in anchorage-independent cell growth control using a library of randomized hairpin ribozymes. Genomics 66:274-283.??:}%, 百拇医药
XU, L., C. K. GLASS, and M. G. ROSENFELD, 1999 Coactivator and corepressor complexes in nuclear receptor function. Curr. Opin. Genet. Dev. 9:140-147.??:}%, 百拇医药
ZHENG, P., J. EASTMAN, S. VANDE POL, and S. W. PIMPLIKAR, 1998 PAT1, a microtubule-interacting protein, recognizes the basolateral sorting signal of amyloid precursor protein. Proc. Natl. Acad. Sci. USA 95:14745-14750.(Burt Richards Jon Karpilow Christine Dunn Isaac Peterson Andrew Maxfield Ludmilla Zharkikh Majid Abe)