| Nucleic Acids Research | Pages |
Position-independent expression of a human nerve growth factor-luciferase reporter gene cloned on a yeast artificial chromosome vector
Introduction
Materials And Methods
Isolation and analysis of NGF YAC clones
Isolation and physical maps of the cloned NGF YACs
Modification of NGF CEPH764C5
Plasmid-based reporter genes
Mammalian cell culture and transfer of reporter genes
Reporter gene assays
Results
Expression and dexamethasone regulation of reporter genes in L929 cells
Regulation of the NGF gene in L929 cells by other reagents
Discussion
YAC-based reporter genes in mammalian cells
Transcriptional regulation of expression
Posttranscriptional regulation of NGF expression
Outlook and conclusions
References
Position-independent expression of a human nerve growth factor-luciferase reporter gene cloned on a yeast artificial chromosome vector
ABSTRACT
INTRODUCTION
Nerve growth factor (NGF) is the first and best studied member of the neurotrophin gene family. The neurotrophins are small proteins which were identified as factors required for the survival and differentiation of nerve cells in vitro. These proteins play an important role in the development, maintenance and repair from injury in the central nervous system. NGF is active on Schwann cells and sensory, sympathetic and cholinergic neurons. Outside of the nervous system, NGF has a modulating influence on the activity of the immune system with demonstrated effects on differentiation and activity of monocytes, basophils and lymphocytes. NGF-responsive cells express either one or both of the two distinct glycoprotein receptors identified for NGF, a 140 kDa receptor (the trk proto-oncogene products) and a 75 kDa low affinity receptor. Cell types known to synthesize NGF include the neurons, Schwann cells, fibroblasts, astrocytes, smooth muscle cells, CD4+ T-cells, macrophages, mast cells and keratinocytes and in the mouse the submandibular duct epithelium. Recent experiments in vivo support the notion that the survival of neurons both in the aging brain and after axotomy can be enhanced by elevating the concentration of NGF by implanting capsules which release NGF, or by implanting cells transfected with NGF-production genes (1-5). Whereas these experiments support the idea to use locally delivered NGF as a drug, it would however be more advantageous to have non-invasive methods, ideally low molecular weight pharmaceuticals to increase neurotrophin concentrations. Such compounds can be identified with help of neurotrophin-expressing cell lines carrying suitably engineered NGF-reporter genes. The same appoach could be used to identify inhibitors of NGF synthesis, which might be used to treat conditions such as systemic lupus erythematosus, which are characterized by an overproduction of NGF (6).
With conventional reporter-genes the expression levels are often variable and their expression is not necessarily subject to the normal regulatory signals. The major sources of the variation, between different transfected cell clones, are the positive or negative influences from endogenous enhancer- and silencer-like genetic elements in the neighborhood of the integrated transgene. One way to circumvent this problem of position-dependent expression is to use YAC-based reporter genes. YACs (yeast artificial chromosomes) can carry sufficient foreign DNA to include those regions that naturally flank the gene of interest (7). In these flanking regions functional DNA segments for the attachment of chromosomes to the nuclear matrix have been identified called matrix attachment regions (8) or scaffold-associated regions (9-11). These may or may not coincide with locus control regions (9-11), which also are important determinants of the chromatin structure of genetic loci (which may include multiple genes) and of their transcriptional regulation. The presence of such elements probably explains why in cell lines (12) and transgenic mice (13-20) YAC-based transgenes follow precisely the developmental expression program of the native genes.
To obtain a reporter gene which is regulated like native NGF, we have isolated and characterized human YAC clones which carry the entire NGF gene. One of these was converted to a reporter gene by inserting the coding portion of the firefly luciferase into the NGF gene and introducing a neo resistance gene near one of the YAC telomeres. Finally, this modified 800 kb human chromosome segment was introduced into mouse L929 fibroblasts and the regulation of the expression of the luciferase reporter gene was compared with the NGF production by the endogenous NGF gene.
MATERIALS AND METHODS
Isolation and analysis of NGF YAC clones
Genomic DNA plugs from Saccharomyces cerevisiae containing YACs (BioRad kit) were prepared essentially following procedures recommended by the manufacturer. Proteinase K was inactivated using Pefabloc (Boehringer Mannheim) before digestion of the DNA with restriction enzymes and analysis on pulsed field agarose gels (CHEF-DRIII system, BioRad). DNA was transferred to nitrocellulose membranes done following standard Southern blot procedures except that to ensure complete transfer of the DNA in the gel was treated with 0.25 M HCl for 30 min and the capillary transfer extended to 48 h. Random primed labeling of DNA fragments, hybridization, washing and autoradiography were done according to standard methods.
Isolation and physical maps of the cloned NGF YACs
Two YACs containing the human NGF gene, one of 800 kb and the other of 1000 kb, were isolated by screening of pooled DNA of YAC clones of the CEPH human mega-YAC library from Research Genetics using two pairs of oligonucleotides: GAGAGGCCAAAAGCTCCGG with GCATGCCATCCAGCCCC (fragment cloned as pFBY197) and GCTCAGCGTCCGGACCC with GAGTGGGCTAGGGGAGC, spanning the intron-1-exon-2 and exon-2-intron-2 boundaries respectively. A detailed map was constructed using cloned PCR fragments and rescued fragments of the YACs themselves as probes (Fig. 1). The presence of the first exon, which is only 33 bp long, was mapped using a synthetic oligonucleotide corresponding to this sequence as probe to be close to the BsiWI site, which in CEPH754D5 is 60 kb from the Trp1-containing end. Exon-2 and exon-3 lie on either side of the SfiI site which is [sim]180 kb from the Trp1 end of CEPH754D5 and 170 bp from the Trp1 end of CEPH764C5. In pulsed field electrophoresis experiments the distance between the BsiWI and SfiI sites was estimated to be <50 kb, which is in good agreement with the previously published estimate for intron-1 of [sim]40 kb. pFBY195 contains a 548 bp fragment of human genomic DNA obtained by PCR with GCATGGATCCGGGTACCAGTTCTGAGGC and CATGCCCATGGACATTACGCTATGCACCTGG, which was cloned as BamHI-NcoI fragment in pTZ18Nco. The BamHI site is located in the second intron and the NcoI site is at the second methionine of NGF in exon-3. pFBY196 contains a PCR fragment obtained with primers GCATAAGCTTCTCCCAACACCATCAC and GCATCTCGAGTTGTTCTACACTCTGATCACAGC. The HindIII site is located at the initiation codon of human NGF and the XhoI site is downstream in exon-3. pFBY199 contains a genomic fragment obtained by PCR with CCCTCCCTACCTCAACC and GTTGCAATCCAGCAGGC, which hybridizes to the 3[prime] end of exon-3. Two YAC-derived probes were used to characterize the overlap between the two NGF YACs. pFBY256 was obtained by rescuing the pUC/TRP1/CEN-containing end of CEPH754D5 into Escherichia coli. Yeast DNA was cut with NcoI plus XhoI, blunt ended, ligated and transformed into DH5F[prime] cells. A plasmid of 8.5 kb containing 3.2 kb of human DNA was obtained. Fragments were isolated which do not hybridize to the cloning vector sequences and used to probe NGF containing YACs. 1.6 and 1.0 kb EcoRI-SacI fragments and the 2.4 kb EcoRI-PstI fragment hybridized to the human DNA at the pUC/TRP1/CEN end of CEPH754D5 and to CEPH764C5, [sim]80 kb upstream of exon-1. pFBY257 was obtained by rescuing the pUC/TRP1/CEN containing end of CEPH764C5 into E.coli. Yeast DNA was cut with NcoI plus XhoI, blunt ended, ligated and transformed into E.coli DH5F[prime] cells. A 7 kb plasmid containing 1.7 kb of human DNA was obtained. A 0.5 kb EcoRI-XbaI fragment or a 1.3 kb EcoRI-PstI fragment hybridized to the human DNA at the pUC/TRP1/CEN end of CEPH764C5 and to CEPH754D5, [sim]150 kb downstream of exon-3. The cloned DNA was also probed with plasmid ATCC 41030, which contains a 1.45 kb EcoRI fragment of the N-ras gene cloned into pUC12. This gene is closely linked to NGF, being [sim]800 kb away. The N-ras DNA fragment hybridizes to CEPH754D5 (near the URA3 end) but not to CEPH764C5. Both YACs contain all three NGF exons, but because in the smaller of the two YACs, CEPH764C5, the first exon of the NGF gene is further away from the nearest end (180 kb), this clone was primarily used in further experiments, although some of the constructs were made in both YACs as a control.
Figure 1. Map of the human NGF YAC-clones. CEPH754D5 and CEPH764C5 both contain the three NGF exons and the introns of [sim]40 and 6.8 kb respectively. TEL: telomere. Restriction enzymes: B, BstWI; K, KspI; Su, SfuI; S, SfiI; N, NhuI; M, MluI; Nt, NotI; Bs, BssHI.
Modification of NGF CEPH764C5
By homologous recombination in vivo, a variety of plasmids were recombined to generate a YAC that contains a luciferase gene integrated into the NGF gene and a neo gene at one end of the YAC. For these in vivo gene engineering experiments the LEU2 gene of the yeast host was first disrupted by transfection of a LEU2 gene, in which the central BsrGI-DraIII fragment was replaced by the LYS2+ gene, followed by selection for lysine prototrophy. Precise disruption of the LEU2 gene, demonstrated by leucine auxotrophy, was confirmed using Southern blots with LYS2+ and LEU2 probes.
To introduce the luciferase gene into CEPH764C5, the LEU2 disrupted yeast strain was transformed with the HindIII-fragment of pFBY202, a plasmid containing 548 bp of NGF sequence immediately upstream of the initiation codon, fused to luciferase with a two amino acid extension, followed by a LoxP site, the yeast LEU2 gene, another LoxP site and 496 bp of sequence immediately downstream of the third amino acid (Met) of NGF. Following homologous recombination between the NGF DNA on either side of the luciferase-LEU2 a YAC containing a perfect luciferase fusion gene was obtained with a minimum of disruption of the NGF gene. Subsequently, this yeast strain was transformed with pH307-ILV2, an ARS (autonomous replication sequence)-containing sulfometuron resistance plasmid with a CRE recombinase gene transcribed from the GAL1 promoter. This resulted in elimination of the LEU2 gene from the fusion gene by CRE-mediated recombination between the two LoxP sites on either side. The unstable ARS-based plasmid was then eliminated by growing the yeast strain in medium without sulfometuron. The final properties of the CEPH764C5LucNeo NGF-luciferase fusion gene are summarized in Figure 2A.
Figure 2. NGF-luciferase reporter genes. (A) The CEPH764C5LucNeo luciferase reporter gene was created by inserting the DNA coding for luciferase at the third amino acid codon of NGF, a methionine codon. The luciferase sequence ends just behind the AAUAAA polyadenylation sequence at an EcoRI site. Next follows a linker segment containing a loxP recombination site, which at an XhoI site is connected to the second part of the NGF gene. (B) The 50 kb NGF-luciferase reporter gene, composed of the exon-1 and -2 and the engineered luciferase in exon-3, is situated between 550 kb of human genomic DNA upstream and 170 kb downstream (not drawn to scale). To introduce a selection marker for mammalian cells, the TEL-URA3 end of CEPH764C5 (Fig. 1) was replaced with a segment containing TEL-URA3-LEU2-neo by homologous recombination in vivo. The CEPH764C5LucNeo reporter gene construct is shown in comparison with pNGF2, a 7 kb plasmid based reporter gene composed of 2 kb DNA fragment containing the mouse NGF promoter, luciferase followed by splicing and polyadenylation signals from the SV40 early gene. To insert a neo selection marker for mammalian cells plasmid pCEPH754D5 (ATCC 67379) was modified as follows. First a NotI linker was placed at the position of the unique EcoRI site. Next a SalI-XhoI fragment containing a neo gene flanked by the SV40 early promoter a polyadenylation site from the Herpes simplex virus thymidine kinase gene was inserted at the unique SalI site. Lastly another SalI-XhoI fragment with the yeast LEU2 gene was inserted at the SalI site next to the neo gene. A NotI-BamHI fragment of this plasmid was used to transform the yeast strain carrying the CEPH764C5LucNeo NGF-luciferase fusion gene. Following homologous recombination the TEL-URA3 end (Fig. 1) is replaced by a segment containing TEL-URA3-LEU2-neo (CEPH764C5LucNeo, Fig. 2B).
Plasmid-based reporter genes
Plasmid pSV40LucNeo contains a constitutively expressed luciferase gene controlled by the SV40 promoter and enhancer. It was created from pGL2-control (Promega Corp., Madison, WI) by insertion of a BamHI fragment containing a geneticin resistance gene driven by the promoter from the Herpes virus thymidine kinase gene taken from plasmid pCGA20b (21). In plasmid pMuNGFLuc (22), the luciferase gene together with the SV40 small t antigen splicing and polyadenylation site from pGL2-basic (Promega) was placed under the transcriptional control of a 2 kb DNA fragment from the mouse nerve growth factor gene (-1881/+289; GenBank M33683). For selection purposes, it was transfected in combination with pSV2neo (23). Plasmid pMMTVLucNeo was constructed by inserting the luciferase DNA, excised as HindIII fragment from pT3T7luc (Clontech, GenBank # U02437), into HindIII-cut pLKneoBH (24,25). In this plasmid transcription of the luciferase is under control of the dexamethasone-stimulated long-terminal-repeat promoter of the mouse mammary tumor virus (MMTV).
Mammalian cell culture and transfer of reporter genes
Mouse L929 fibroblasts (CCL1) obtained from the American Type Culture Collection (Rockville, MD) were subcultured 1/10 twice weekly in Dulbecco's minimal essential medium (DMEM) medium supplemented with 10% fetal calf serum (FCS) (all from Life Sciences Inc.). Cells were treated with water-soluble cyclcodextrin-complexed dexamethasone (Sigma Corp., St Louis), recombinant bFGF (Life Technologies Inc., Gaithersburg, MD), interleukin-1 (IL-1, Ciba-Geigy Ltd, Basel), Okadaic acid (LC Laboratories), dBcAMP (Sigma) or transforming growth factor-[beta]-3 (TGF-[beta]-3, Ciba-Geigy Ltd).
The CEPH764C5LucNeo reporter genes were transferred to mammalian cells by spheroplast fusion (26). Yeast cells of the AB1380 strain containing the engineered YAC reporter gene (OD660 of 37.5, 15 million cells/µl 0.8 M sorbitol) were shaken at 110 r.p.m. with 60 µg/ml zymolase for 15 min at 30°C. The OD660 decreased by 80%. The spheroplasts were collected by centrifugation, washed twice in 0.8 M sorbitol. Final concentration was adjusted to OD660 2.0. L929-cells were trypsinized at 75% confluence, washed twice in serum-free DMEM and resuspended at a final concentration of 2 million cells/ml.
Fusion with the spheroplasts was achieved by collecting 2 million L929-cells on the bottom of a 15 ml polystyrene conical centrifugation tube. The supernatant was aspirated and 1 ml of the yeast spheroplasts suspension was layered on top of the cell pellet. The ratio of spheroplasts to mammalian cells was 50:1. After centrifugation (10 min, 600 g) cell pellets were resuspended in 50 µl serum-free DMEM. Five hundred µl of a 50% polyethylene glycol 1500 solution (Boehringer Mannheim) and 3 µl 1 M CaCl2 were mixed with cell suspension by gentle inversion. The cells were incubated at room temperature for 2 min, then diluted with 5 ml of serum-free DMEM and after an additional 30 min incubation at room temperature, the cells were collected by centrifugation and plated out in complete medium. The next day, floating cells were removed and fresh complete medium was added to the adherent cells. 48-72 h after the fusion, the medium was replaced by complete medium with 0.6 mg/ml Geneticin. 50-100 Geneticin-resistant colonies were observed in the third week after the fusion and 24 colonies were isolated in the fourth week. Plasmid-based reporter genes were transfected using the cationic lipid Lipofectin (GIBCO/BRL) as previously described (27).
Reporter gene assays
Luciferase was assayed using reagents and protocols from Promega Corp., Madison, WI. Aliquots (0.02 ml) of cell extract (the progeny of 800 000 cells seeded are lysed in 1 ml lysis reagent, i.e., at the time of lysis [sim]2.4 x 106 cells/ml) were transferred to standard luminometer 96-well assay plates (Microlite-2, Dynatech # 011-010-7417). To 0.02 ml lysis reagent 0.1 ml of luciferase substrate solution was injected immediately prior to the light measurement. The luminometer (Dynatech ML3000) was set in enhanced flash mode, with a delay time (after substrate injection) of 2 s and an integration time of 10 s. Output is given in relative light units (RLU). Secreted mouse NGF was detected in an ELISA assay using anti-mouse-(2.5S) nerve growth factor (Boehringer # 1008 226) as coating antibody and as developing antibody a conjugate of anti-mouse-(2.5S) nerve growth factor and [beta]-galactosidase (Boehringer # 1008 234). The latter was subsequently detected using 4-nitrophenyl-[beta]-d-galactopyranosid as substrate.
RESULTS
The CEPH764C5LucNeo reporter genes were transferred to mammalian cells using the spheroplast fusion method (Table 1). About 50-100 neo colonies containing the CEPH764C5LucNeo NGF reporter gene were obtained in two consecutive experiments. The state of CEPH764C5LucNeo in the transformed cell was not investigated, but the two most studied clones, YAC-L6 and YAC-L8, continued to express luciferase after over 20 passages without selection. These two CEPH764C5LucNeo-transformed cell lines were also stable during supertransfection experiments with a plasmid containing the gpt selection marker. This stability suggests that the CEPH764C5LucNeo is associated with the centromere of one of the L929 cell chromosomes. From a practical point of view, both cell lines are also sufficiently stable to be used in a random drug screening operation.
Parallel to the CEPH764C5LucNeo-containing cell lines, several plasmid-transformed cell lines were constructed. This was done with two purposes in mind: first, to obtain cell lines to check the specificity of NGF-regulating compounds, and second, to be able to compare the range of responsiveness of these luciferase reporter genes in different cell lines (thought to be a function of their chromosomal integration site) with that of the different CEPH764C5LucNeo-containing clones. Relative to the number of L929 cells transfected, the efficiency of plasmid transfection 100-200 more neo colonies were obtained than with CEPH764C5LucNeo spheroplast fusion. Several hundred colonies were obtained/µg plasmid with each of the plasmids.
Expression and dexamethasone regulation of reporter genes in L929 cells
Luciferase expression was tested with and without 1 µM of the synthetic glucocorticoid hormone dexamethasone (DEX), which is known to inhibit NGF production in L929 cells (28-30). The absolute level of the luciferase expression in the different CEPH764C5LucNeo-transfected cell lines varied only over a extremely narrow range (Fig. 3A and B and Table 1). The 3-fold variation observed is even further reduced if corrections are made for slight differences in cell numbers (not shown). However, the basal level of expression of the YAC-based reporter gene was very low compared to the lowest expression level observed with the plasmid-based reporter genes (Fig. 3 and Table 1). The level of luciferase mRNA transcribed from the CEPH764C5LucNeo gene assayed on northern blots was similarly low compared to the level of endogenous mouse NGF mRNA (Fig. 4).
The luciferase expression of pMuNGFLuc-transfected cells varied over four log units (Fig. 3C). The inhibition by DEX was varied much more than among the CEPH764C5LucNeo-derived cell lines, than among those transfected with pMuNGFLuc (Table 1), but, like with the YAC-based reporter gene, the level of DEX inhibition of the plasmid-derived reporter genes was not affected by the widely different basal expression levels, as shown by a linear correlation between the DEX-inhibited and control luciferase expression levels.
A wide variation of luciferase expression was also observed in pMMTVLucNeo-transfected cell lines (Fig. 3D). In this case, DEX stimulation typically increased in proportion to the basal level of luciferase expression being 3-fold in the clones expressing the smallest amount of luciferase to 12-fold in the high-expressing clones. There were a few clones that did not fit this pattern, but for over 90% the inducibility appeared to rise by 3.6 ± 0.6 units for every log unit increase of the basal luciferase expression level (Fig. 3E) or the induced expression level (Fig. 3F).
Of 24 pSV40LucNeo-transfected cell lines, six did not express significant luciferase activity (RLU < 0.005), four expressed large amounts of luciferase and the rest expressed intermediate amounts. These results (Table 1) are typical for this plasmid. Similar results have also been obtained with mink lung cells (Mv1Lu), Chinese hamster ovary cells, and human glioblastoma cells (U373) (not shown). Luciferase expression from the SV40 promoter/enhancer in pSV40LucNeo was not affected by 1 µM dexamethasone.
Table 1.
| Reporter gene | N | Luciferase | Dexamethasone inhibition/stimulation | |
| RLU ± SD (SD as %) | fold ± SD | % of control ± SD (SD as %) | ||
| CEPH764C5LucNeo | 36 | 0.018 ± 0.008 (±43%) | 0.47 ± 0.07 | 47.8 ± 7.4 (±14%) |
| pMuNGFLuc | 42 | 3.2 ± 5.8 (±180%) | 0.81 ± 0.31 | 84. 3 ± 32.6 (±127%) |
| pMMTVLucNeoa | 24 | 0.04 ± 0.09 (±208%) | 5.25 ± 3.82a | 527 ± 382 (± 72%)a |
| pSV40LucNeo | 18 | 0.11 ± 0.21 (±189%) | - | - |
Figure 3. Expression of transfected reporter genes in mammalian cells. (A) Linear plot with 95% confidence boundaries (dotted lines) of the luciferase expression levels in 18 cell lines containing the CEPH764C5LucNeo reporter gene after 48 h treatment with and without 10-6 M dexamethasone (DEX), n = 34 R = 0.945 F = 2666.3 P < 0.0001. (B) Double logarithmic plot of luciferase expression levels in 24 cell lines stably cotransfected with pMuNGFLuc and pSV2neo, n = 42 R = 0.973 F = 713.2 P < 0.0001. (C) Double logarithmic plot of expression levels in 24 cell cell lines stably transfected with pMMTVLucNeo, n = 23 R = 0.989 F = 918.1 P < 0.0001. (D) DEX stimulation of the luciferase expression in pMMTVLucNeo-transfected cell lines as a function of the expression levels with DEX treatment. Dose-response curves were established with two cell lines each containing either pMMTVLucNeo (induced by DEX), pMuNGFLuc and CEPH764C5LucNeo (inhibited by DEX). Results are expressed in units of measurement (Fig. 5A-C) or as percentile values (Fig. 5D-F). Luciferase content of the cells and the amount of NGF in the medium were determined 40 h after DEX addition. DEX inhibited expression of the luciferase reporter genes with an EC50 of [sim]10 nM, but NGF secretion already at 1-2 nM. The reporter cell lines were also supertransfected with pGRgpt, a plasmid with a gpt selection marker, which expresses the human glucocorticoid receptor (31). However, overexpression of the glucocorticoid receptor did not affect expression of NGF or the NGF reporter genes. As also the expression of the luciferase from the MMTV promoter was not affected, the concentration of receptor is probably not limiting for either the downregulation of the NGF promoter or for the upregulation of MMTV promoter. Figure 4. Northern blot. 5 x 105 CEPH764C5LucNeo transfected L929 cells were cultivated in serum-free DMEM (-) or stimulated with 10% FCS (+) for 8 h. Poly(A)+ RNA was extracted, blotted and hybridized with a digitonin-labeled cRNA probe for endogenous NGF and luciferase reporter mRNA expression. The size of the mouse NGF mRNA is 1.3 kb (28). The observed size of the luciferase-fusion mRNA is estimated at 2.4 kb (expected size 3.2 kb). Figure 5. Dexamethasone dose-response with different reporter genes in stably transformed L929 cell lines. Dose-response relationship was determined with L929 cell lines stably transformed with each of the different reporter genes. Luciferase content of the cells and the concentration of mouse NGF secreted in the culture medium were measured after 48 h DEX treatment. Results are given as measured activities (A-C) or calculated as percentage of the control (without DEX). 
Regulation of the NGF gene in L929 cells by other reagents
A series of agents which have been described to affect NGF production in L929 fibroblasts were tested (Fig. 6). In this series of experiments, in which NGF and luciferase were measured 8 or 16 h after reagent addition, the effect of dexamethasone on luciferase expression was confirmed, although the effect on NGF secretion was not as dramatic as seen after 48 h (Fig. 5). Addition of dibutiryl-cyclic AMP stimulated NGF secretion much more than NGF-luciferase expression (Fig. 6), suggesting that NGF regulation by cAMP (28,32) is primarily posttranscriptional. Okadaic acid stimulated NGF secretion, but the stimulation of luciferase expression was to small to be statistically significant. TGF-[beta] inhibited NGF secretion by 16-24%, but stimulated the luciferase expression somewhat, although the latter result was not statistically significant. Considering the inhibition by TGF-[beta], it should be considered that TGF-[beta] is a mitogen for L929 cells (33). As a consequence, the net effect of TGF-[beta] cannot be studied without taking cell proliferation into consideration. Phorbol esters have previously been reported to stimulate NGF mRNA levels in L929 cells (28,34). In this series of experiments in medium with 5% FCS significant stimulation of NGF secretion by phorbol 12-myristate 13-acetate (PMA) was observed in one of the four experiments in the CEPH764C5LucNeo-transformed cells. Likewise, significant stimulation of luciferase expression was observed in one of four experiments, however not in the same experiment as the stimulation of NGF secretion (Fig. 6). The presence of serum in the medium is important because serum stimulation occurs in part, like PMA, through the protein kinase C pathway. Without FCS in the medium, PMA consistently stimulated NGF synthesis several-fold and serum-containing medium by itself stimulated synthesis of both the endogenous NGF mRNA as well as the luciferase fusion mRNA (Fig. 4). Interleukin 1 has been reported to stimulate NGF synthesis in skin fibroblasts (35), however no significant effect was found in our L929 experiments either on NGF secretion or luciferase expression.
Figure 6. Regulation of NGF-reporter gene expression. Secreted mouse NGF (A) and luciferase (B) were tested 8 h after compound addition (solid bars) or after 24 h (three other experiments) with clone YAC-L6. Compounds tested were: 1, 10 ng/ml bFGF; 2, 1 mM dibutyryl-cAMP; 3, 100 ng/ml PMA; 4, 0.01 mM okadaic acid; 5, 10 ng/ml TGF-[beta]; 6, 1 mM dexamethasone. Similar results were obtained with clone YAC-L8. After addition of basic fibroblast growth factor (bFGF) both NGF-secretion and expression of NGF-YAC-reporter gene were upregulated considerably, although after 6 h, only the reporter gene was stimulated (Fig. 6). Presumably, the effects on accumulation of NGF in the medium is delayed in time with respect to the effects at the level of transcription. The dose-response relationship observed in the response of the luciferase reporter gene and the endogenous mouse NGF gene is very similar, both at 16 h and 20 h after addition of bFGF (Fig. 7). 
DISCUSSION
YAC-based reporter genes in mammalian cells
Of the three techniques used to transfer YACs from yeast to mammalian cells, PEG-mediated spheroplast-mammalian cell fusion, microinjection and lipofection, we have chosen spheroplast-cell fusion, because it had been used successfully with a number of rodent cell types including mouse L929 fibroblasts (36-42). Although spheroplast-cell fusion reportedly has worked with at least one human cell line (36), in our hands, this method has only worked with the mouse L929-cells and not with human U373 glioblastoma or NT-2 neuronal precursor cells.
Figure 7. Regulation of the CEPH764C5LucNeo luciferase by basic FGF. Luciferase content of CEPH764C5LucNeo-transformed L929 cells (clone L6) and mouse NGF secreted in the culture medium were measured after bFGF treatment during 16 or 24 h. The most dramatic difference between the YAC- and plasmid-based reporter genes has been the narrow range of variation in the level of expression of the YAC reporter gene. Whereas the expression from the plasmid-derived luciferase gene varied up to 10 000-fold between different clones, this variation was <3-fold in the YAC-derived cell lines (Fig. 3B). This suggests that this gene expression is truly position-independent. The spread in the results was sufficiently narrow that a relevant estimate of degree of inhibition by dexamethasone could be obtained from the average performance of the different CEPH764C5LucNeo-transfected clones (Table 1). This was clearly not the case with the plasmid-based reporter gene, even though two thirds of these clones were inhibited by dexamethasone by >20%. The basal level of this position-independent luciferase expression was much lower than the lowest expression observed with the plasmid-based NGF reporter gene. This may be due to the presence of silencer-type genetic elements in the YAC gene constructs or due to a lower stability of the luciferase mRNA produced by the YAC-based gene. 
Transcriptional regulation of expression
The dose-response relationships observed in the response of the luciferase reporter gene and the endogenous mouse NGF gene are very similar, both at 16 h and 20 h after addition of bFGF (Fig. 7). Therefore it appears that the reporter gene is a good indicator of the stimulatory pathway from the FGF receptor(s) to the NGF-promoter (43).
Dexamethasone downregulates in L929 fibroblasts both the production of NGF protein and the luciferase expression of the NGF-reporter gene. However, five times more dexamethasone is needed to downregulate the reporter gene than to suppress the secretion of endogenous mouse NGF. The EC50 for the NGF and MMTV luciferase reporter genes is similar, suggesting that both are due to the same mechanism, presumably the binding of dexamethasone to the glucocorticoid receptor, which as a transcription factor exerts a direct effect on gene expression. It has been shown that, in the case of regulation of NGF expression by IL-1 and cAMP, this is in part due to alterations in the stability of the NGF mRNA (44-47). Possibly, dexamethasone affects the secretion of mouse NGF in addition to its effects on transcription also through such an indirect pathway, leading to a different dose-response relationship with respect to production of NGF protein. A similar situation exists in the case of inhibition of interleukin-2 expression in T lymphocytes. Also in this case production of interleukin-2 protein is much more suppressed than transcription of the interleukin-2 gene, presumably because under the influence of dexamethasone the half-life of the interleukin-2 mRNA is reduced (48).
Posttranscriptional regulation of NGF expression
Furthermore, some of the effects may be related to the posttranscriptional component in the regulation of NGF expression. The situation is complicated somewhat, because-contrary to our initial intention-the luciferase cDNA segment, which was inserted into the human NGF gene, included the polyadenylation signal of the luciferase gene (Fig. 2). It is therefore expected that the mRNA transcript coming from the NGF-promoter terminates at this point and does not include the 3[prime] end of the NGF gene. Indeed the size of the recombinant luciferase mRNA on northern blots (Fig. 4) would confirm this hypothesis, as does the fact that, in contrast to native human NGF mRNA, the mRNA of the reporter gene did not hybridize strongly with mouse NGF cDNA.
The absence or presence of the 3[prime] untranslated region of the NGF mRNA in the luciferase fusion mRNA may have important consequences for the regulation of expression, because of the occurrence of multiple AU-rich elements (AREs) in the 3[prime] UTR of the NGF mRNA. In mouse, but in human NGF mRNA the Ares include the most potent variant UUAUUUAUU (49), although direct involvement of the AREs in the regulation of mRNA abundance as has been shown for numerous oncogenes and cytokine mRNAs (50,51), this remains to be proven in the case of NGF. However, several lines of circumstantial evidence suggests that this might be the case. Multiple signal transduction pathways, which affect NGF expression, are known to regulate other mRNA species through the AU-binding protein, which when bound to the AREs increases mRNA stability. For example regulation of the stability of NGF-mRNA by IL-1 has been demonstrated by two groups (45,52) and the ARE-binding factors is probably the main signal transduction pathway activated by IL-1 as almost all IL-1-responsive genes contain AU-rich sequences in the 3[prime] untranslated region (53). In this respect, it is noteworthy that, in contrast to secretion of the endogenous mouse NGF, our transfected NGF reporter genes do not respond to IL-1. Activation of the AU-binding factor by phorbol esters has been demonstrated for intracellular adhesion molecule-1 (54) and GLUT-1 mRNA (55). It has been shown that some of the effects of phorbol esters occur by stabilization of the NGF mRNA (34). Therefore, it is not unlikely that this also occurs for NGF though the AU-binding factor. Glut-1 mRNA is also stabilized by cAMP and okadaic acid (55). The latter compound is known to stabilize NGF mRNA in astrocytes (56). It is not unlikely that dexamethasone in part decreases NGF expression in L929 cells by lowering the concentration of the ARE-binding factor as this has been demonstrated for [beta]-interferon mRNA in these cells (57).
Outlook and conclusions
The NGF example demonstrates convincingly the feasibility of the YAC-reporter gene approach.. Even if not all questions about NGF regulation have been answered, it is clear that the CEPH764C5LucNeo-transformed L929 fibroblast lines seem to respond in highly reproducible physiological meaningful manner. Therefore, they could in principle be used to discover compounds which regulate transcription of the NGF gene. If, as preliminary evidence suggests, the CEPH764C5LucNeo luciferase mRNA does not contain the NGF 3[prime] untranslated region, this might turn out to be an advantage. It would avoid detection of compounds regulating NGF through the ARE-binding protein pathway, which would probably not be NGF specific. The CEPH764C5LucNeo L929 cell lines could also present an important model system to dissect the transcriptional and posttranscriptional components of the NGF regulation in L929 cells.
ACKNOWLEDGEMENTS
The authors want to express their gratitude to Dr Roger Aeschenbacher for plasmid pSV40LucNEO and to Dr Christine Gandor for pMMTVLucNeo.
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 24 Mar 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||