Molecular and Cellular Endocrinology, 69 (1990) 93-99 Elsevier Scientific Publishers Ireland. Ltd.
MOLCEL
93
02231
Human growth hormone gene expression in rat but not human non-pituitary cells after stable gene transfer Barbara E. Nickel, Mark N. Nachtigal,
Margaret
E. Klassen and Peter A. Cattini
Department of Physiology, University of Manitoba, Winnipeg, Manitoba R3E 0 W3, Canada (Received
Key words: Growth
hormone;
Tissue-specific
23 August
expression;
1989; accepted
22 November
1989)
Gene transfer
Summary Tissue-specific expression of the rat growth hormone (rGH) gene requires binding of a pituitary-specific factor. Binding of this factor has been used to explain tissue-specific expression of the human growth hormone (hGH-N) gene in transfected rat pituitary (GC) tumour cells. Neither rat fibroblast (R2) nor human cervical carcinoma (HeLa) cells contain the rat pituitary-specific factor. Thus, no expression of hGH-N or rGH would be expected in these cells. R2 cell lines containing stably integrated hGH-N or rGH genes were generated. Expression of hGH-N but not rGH was detected. By contrast, stably transfected HeLa cells did not express the endogenous or transfected hGH-N genes. However, an hGH-N transcript was detected when hGH-N gene expression was directed by a viral promoter. This suggests that the block in expression occurs at the level of transcription and not mRNA stability. Hybrid genes containing 496 base pairs (bp) of hGH-N or 234 bp of rGH 5’-flanking DNA, including promoter sequences, fused to the bacterial gene coding for chloramphenicol acetyltransferase were used to stably transfect R2 cells. The hybrid hGH-N gene was more active than a promoterless construction in these cells. By contrast, the hybrid rGH gene was not. These data suggest that the hGH-N gene can be activated by rat transcription factors other than those found in pituitary cells.
Introduction Human pituitary growth hormone (hGH-N) and chorionic somatomamrnotropin (hCS) are two members of the human growth hormone hGH family (Eberhardt et al., 1988). The genes coding for these hormones share greater than 90% nucleotide sequence homology yet production of hGH-N is limited to the somatotrophs of the
Address for correspondence: Barbara E. Nickel, Department of Physiology, University of Manitoba, 770 Bannatyne Avenue, Winnipeg, Manitoba R3E 0W3, Canada. 0303-7207/90/$03.50
0 1990 Elsevier Scientific
Publishers
Ireland,
anterior pituitary and hCS to the syncytiotrophoblast of the placenta (Miller and Eberhardt, 1983). The mechanism by which pituitary- and placenta-specific members of the hGH gene family are differentially expressed in a tissue-specific manner is still unclear (Chen et al., 1989; Nachtigal et al., 1989). Preferential expression of the rat growth hormone (rGH) gene and the hGH-N gene is obtained in rat anterior pituitary tumour (GC) cells compared to non-pituitary cells after gene transfer (Cattini et al., 1986). This activity has been explained by the presence of a pituitary-specific factor variously called GHF-1, pit 1 GCl or 2 in these cells (Catanzaro et al., Ltd
94
1987; Lefevre et al., 1987; West et al.. 1987). The absence of this factor in non-pituitary cells has led to the conclusion that binding of GHF-1 results in tissue-specific expression of hGH-N as well as rGH. However, two placenta-specific members of the hGH gene family, hCS and placental growth hormone variant (hGH-V), are also expressed in GC cells after gene transfer (Nachtigal et al., 1989; Nickel et al., 1990). Binding of GHF-1 to the hCS-A or hCS-B genes which are both able to code for hCS has been demonstrated (Lemaigre et al., 1989; Nachtigal et al., 1989) and, in the case of hCS-A, is responsible for promoter activity in transfected GC cells (Nachtigal et al., 1989). A possible explanation to help reconcile these data is that tissue-specific expression of the hGH gene family might be accomplished through differential repression. There is evidence for repressor activity in the 5’flanking DNA of the hGH-N (Peritz et al., 1988) and hCS-A genes (Nachtigal et al., 1989). The hGH/hCS genes would be activated by factors such as GHF-1 or their non-pituitary cell equivalents, and silenced by pituitaryor placenta-specific repressors. As a consequence, members of the hGH gene family might be activated by factors other than GHF-1, if the appropriate repressors are absent. To investigate the role of repression and the possibility of additional activating factors, we examined hGH-N gene expression in non-pituitary rat and human cells after gene transfer. Our results suggest that the hGH-N gene can be activated by rat transcription factors other than GHF-1, and suggest that efficient silencing of the hGH-N gene occurs in human but not rat nonpituitary cells. These data are discussed in relation to current ideas on tissue-specific expression of the hGH gene family.
(-496hGH-Np.car). hCS-A (-496hCS-Ap.cat) or 234 bp of rGH 5’-flanking DNA ( - 234rGHp.cat) fused upstream of the bacterial gene coding for chloramphenicol acetyltransferase (CAT). All constructions were made using protocols described in Maniatis et al. (1982) unless otherwise stated. The plasmid containing both the hCS-A and neomycin resistance gene (pneohCS-A) was constructed as previously described for pneohGHN but using the hCS-A gene in place of the hGH-N gene (Cattini et al., 1986a). The truncated rGHm gene was generated by deleting 5’-flanking sequences -235 (BgfII) to - 1753 (EcoRI) from prGHm (Cattini et al., 1986a). A hybrid gene was made containing the herpes-simplex thymidine kinase (TIS) promoter (p.) fused upstream of the hGH-N structural gene sequences. The plasmid phGH-N (Cattini et al., 1986a) was cut with BamHI to remove the 5’-flanking and promoter This fragment was replaced with a sequences. BamHI/Bg/II promoter fragment of TK containing nucleotides - 109 to + 53 (Cattini et al., 1986b) to generate - 109TKp.hGH-N.
Cell culture and gene transfer Human cervical carcinoma (HeLa), rat fibroblast (R2) and anterior pituitary tumour (GC) cells were grown in monolayer culture in Dulbecco’s modified Eagle’s medium supplemented with 2.5% fetal calf serum and 2.5% calf serum. For stable gene transfer, 2-3 X lo6 cells were plated per 100 mm culture dish and transfections were done by calcium phosphate-DNA precipitation as described previously (Cattini et al., 1988). Resistant clones were selected with 0.5 mg/ml (active) G-418 (Sigma) and a number greater than 50 was designated as a pool.
DNA isolation and blotting Materials and methods
Materials Many of the gene or hybrid gene fragments used in this study have been described previously (Cattini et al., 1986a, b; Nachtigal et al., 1989). These include hGH-N, hCS-A, modified rGH (rGHm), the gene coding for neomycin resistance (neodX), neohGH-N, neorGHm, hybrid genes containing 496 base pairs (bp) of hGH-N
Medium was aspirated from the culture dish and the cells washed with 5 ml of phosphatebuffered saline (PBS). The cells were scraped and transferred to a 15 ml tube in a further 5 ml of PBS. The cells were pelleted by centrifugation, and resuspended in 1 ml of PBS before transfer to a 1.5 ml microfuge tube. The suspension was spun for 10 s to pellet the cells and the PBS was aspirated. The cells were lysed by adding 1 ml of 10 mM Ttis-HCl pH 7.6,150 mM sodium chloride
95
(NaCl), 2 mM magnesium chloride (MgCl,) and 0.5% NP40 with incubation on ice for 5 min. The nuclei were spun for 20 s in a microfuge and the supernatant was discarded. The nuclear pellet was resuspended in 0.5 ml of 15 mM Tris-HCI pH 7.5, 160 mM NaCI, 3 mM MgCl,, 10 mM EDTA and 2.2% Na-sarcosyl and then 25 ~1 of 10 mg/ml pronase was added and the mixture incubated overnight at 37” C. The digest was then treated with 10 mg/ml ribonuclease for 3 h followed by 25 ~1 of 10 mg/ml pronase with overnight incubation at 37’C. The protein was removed from the digest by repeated phenol : chloroform steps until the interface was clear (at least three) and the DNA pr~pitated with isopropanol (Maniatis et al., 1982). The DNA was resuspended in 10 mM Tris-HCI pH 7.5, 1 mM EDTA. DNA (5 pg) was digested with restriction endonucleases and then fractionated in 1% agarose gels, transferred to nitrocellulose and assessed with radiolabelled probes according to established protocols (Maniatis et al., 1982). RNA isolation and blotting Total cytoplasmic RNA was isolated by a modification of the procedure of Chomczynski and Sac&i (1987) as described elsewhere (Klassen et al., 1989). Total cytoplasmic RNA was denatured with formaldehyde and fractionated by electrophoresis through a 1.5% agarose gel (Maniatis et al., 1982). The RNA was blotted onto nitrocellulose and probed with 32P-radiolabelled hGH cDNA (Martial et al., 1979), rGH cDNA (Seeburg et al., 1977) or B-actin (kindly provided by Dr. S. deToledo) at 42°C in the presence of 50% formamide. Blots were wahsed 3 times for 15 min in 0.1 x SSC (20 x SSC; 3 M NaCl, 0.3 M sodium citrate) with 0.1% sodium dodecyl sulphate at 65OC and then assessed by autora~o~aphy. To remove the previous probe, the nitrocellulose was washed 3 times with boiling water before subsequent prehybridization and hybridization steps. CA T assays CAT assays were carried out as previously described (Nachtigal et al., 1989) and activity was measured using two-phase fluor diffusion assay (Neumann et al., 1987). Quantitative values for CAT activity were determined by regression anal-
ysis to give counts per minute, per minute, per milligram (cpm/minfmg) of cell protein extract. Results
Expression of hGH-N, KS-A, rGHm and -234rGHm in stably transfected non-pituitary R2 cells Modified rGH genes, rGHm, neorGHm and - 234rGHm, as well as hGH-N, neohGH-N, hCSA and neohCS-A were stably introduced into R2 cells to generate the cell lines R2[rGHm], RZ[neorGHm], R2[ - 234rGHm], R2[hGH-N], R2[neohGH-N], R2[hCS-A] and R2[neohCS-A]. The modification in the rGH genes was made by introducing a BamHI decanucleotide linker into the 5 ‘-untranslated region. This modification allows us to distinguish between the transfected and endogenous rGH gene. We have shown previously that rGHm is expressed along with the endogenous rGH gene in GC cells after gene transfer (Cattini et al., 1986a). Analysis of genomic DNA from the cell lines generated was done to determine that successful integration of the transfected genes had occurred. Genomic DNA from each cell line was isolated and cut with restriction endonucleases which identify the transfected gene; EcoRI or PuuII for hGH-N and hCS-A genes, and Hind111 compared to Hind111 with BarnHI for the endogenous and transfected rGH genes. The DNA was resolved by electrophoresis, transferred to nitrocellulose, probed with radiolabelled hGH cDNA for the human genes and rGH cDNA for the rGH genes, and assessed by autoradiography (Figs. 1 and 2). The integration of hGH-N or hCS-A in the R2 genome is indicated by hybridization of DNA to the hGH cDNA probe compared to untransfect~ R2 DNA, and the presence of approp~ately sized fragments (Fig. 1). The integration of the modified rGH genes is indicated by the presence of additional bands in the transfected compared to the untransfected cell DNA that hybridizes to rGH cDNA. Unlike the endogenous rGH gene DNA which contains no BamHI site within its structural sequences, the bands corresponding to rGHm or - 234rGHm are cleaved by BamHI (Fig, 2). Total cytoplasmic RNA isolated from R2 cell lines containing hGH-N, hCS-A, rGHm and
96
-234rGH, as well as untransfected R2 cells was assessed by RNA blotting. RNA was separated by transferred to nitrocellulose, electrophoresis, probed sequentially with radiolabelled hGH cDNA, rGH cDNA and P-actin, and each was visualized by autoradiography (Fig. 3). RNA from untransfected GC cells and GC cells transfected with the hGH-N gene was also analyzed as controls (Fig. 3). Both hGH-N and hCS-A gene expression was observed in transfected R2 cells. By contrast, neither rGHm, - 234rGH or endogenous rGH mRNA was detected (Fig. 3). Stability of hGH-N transcripts in stably transfected HeLu cells Human cervical carcinoma (HeLa) cells were stably transfected with the hGH-N gene and four separate pools (HeLa[hGH-N]l-4) generated. No
abcdefghij Fig. 1. Detection of hGH-N or hCS-A genes in R2 cells by DNA blotting. DNA (5 pg) isolated from untransfected R2 cells (a,b) or R2 cells stably transfected with hGH-N (cd), hCS-A (e,f), neohGH-N (gh) and neohCS-A genes (ij) was cut with EcoRI (a,c,e,g,i) or PuuII (b,d,f,hj). The DNA was fractionated in a 1% agarose gel, transferred to nitrocellulose, probed with radiolabelled hGH-N cDNA and assessed by autoradiography. No hybridization to untransfected R2 DNA is observed (a,b), and bands corresponding to 2.6 kb hGH-N (c,g) and 2.9 kb hCS-A (e,i) EcoRI fragments, respectively, are seen.
abcdef Fig. 2. Detection of transfected and endogenous rGH genes in R2 cells by DNA blotting. DNA (5 pg) from untransfected R2 cells (a,b) and R2 cells transfected with the rGHm (cd) and -234rGHm genes (e,f) was cut with Hind111 (a,c,e) and HindIII/BumHI (b,d,f). The DNA was resolved in a 1% agarose gel, transferred to nitrocellulose and probed with a radiolabelled rGH cDNA and assessed by autoradiography. Additional bands are seen in the DNA from transfected cells which are cut by BarnHI; BamHI is the site used to make the modification in the rGH gene structural sequences.
endogenous or transfected hGH-N mRNA was detected in 0.1 mg of total cytoplasmic RNA from any of these lines by RNA blotting analysis, even after prolonged autoradiographic exposure (data not shown). This could be due to an undetectable level of hGH-N transcription or instability of hGH-N transcripts in HeLa cells. To determine which of these possibilities are most probable, a hybrid gene( - 109TKp.hGH-N) was constructed in which the hGH-N gene 5’-flanking DNA, including the endogenous promoter, was replaced with the herpes-simplex thymidine kinase promoter( - 109TKp). The - 109TKp.hGH-N gene was stably introduced into HeLa cells and expression assessed by RNA blotting analysis. RNA from HeLa[hGH-N]l and untransfected HeLa cells were included as controls (Fig. 4). A - 109TKp.hGH-N transcript was detected in HeLa cells suggesting that hGH-N mRNA would be stable in these cells if produced. These data suggest that the absence of endogenous or transfected hGH-N mRNA in HeLa cells reflects an effect at the level of transcription and not mRNA stability.
97
28s
18s
a
bc
Fig. 4. Expression of hGH-N in HeLa cells by RNA blotting. Total cytoplasmic RNA was isolated from HeLa cells transfected with (a) - 109TKp.hGH-N or (b) hGH-N (pool 1) and (c) untransfected HeLa cells. RNA (0.1 mg) was resolved in a formaldehyde-agarose gel, transferred to nitrocellulose and probed with radiolabelled hGH-N cDNA, and assessed by autoradiography.
a
bcdefghijk
Fig. 3. Expression of transfected hGH-N, hCS-B and rGH in R2 cells detected by RNA blotting. Total cytoplasmic RNA was isolated from (a) GC[hGH-N], (b) GC, (c,f,k) R2, (d) RZ[hGH-N], (e) RZ[hCS-A], (g) RZ[neohGH-N], (h) IU[neohCS-A], (i) IU[rGHm] and (i) R2[ -234rGHml cells. RNA (20 pg) was fractionated in a formaldehyde-agarose gel, transferred to nitrocellulose and probed with radiolabelled hGH-N, rGH and B-actin cDNAs as indicated, and assessed by autoradiography.
The role of proximal hGH-N and hCS-A S’-jlanking sequences on gene repression Hybrid cat genes were made containing 496 bp of hGH-N, hCS-A or 234 bp of rGH 5’-flanking sequences fused upstream of the bacterial gene (cat) coding for chloramphenicol acetyltransferase (CAT). These genes are expressed efficiently in GC cells after gene transfer (Cattini et al., 1986; Cattini and Eberhardt, 1987). The same hybrid cat genes were used to stably transfect B2 cells, and expression was assessed by measuring CAT activity in a two-phase fluor diffusion assay. A cell line containing a promoterless (-p) cat gene was also
made (B2[ -p.cat] cells) and assayed to give an indication of the level for random initiation of transcription (Table 1). Hybrid hGH-N and hCS-A genes were more active than the -p.cat gene in R2 cells. By comparison, the -234rGHp.cat gene was not expressed. TABLE
1
EXPRESSION OF HYBRID cut GENES DIRECTED hGH-N, hCS-A OR rGH 5’-FLANKING SEQUENCES STABLY TRANSFECTED R2 CELLS Cell line
Gene
cpm/min/mg
R2 R2[ R2[ R2[ R2[
- p.caf - 234rGHp. cur -496hGH-Np.cur -496hCSAp.cat
77.1+ 1os.9+ 94.5 + 283.4+ 347.2 +
-p.car] - 234rGHp. car] - 496hGH-Np.car] -496hCS-Ap.cat]
3.6 4.4 2.4 2.1 1.3
BY IN
(4) (3) (3) (4) (3)
CAT activity was assessed in untransfected R2 cells and R2 cells stably transfected with a promoterless cur gene ( - p. cut), or hybrid car genes containing 234 or 496 bp of ffiH or hGH-N and hCS-A 5’-flanking DNA, respectively. Values for expression are given as cpm/min/mg cell extract plus or minus standard error from the mean. The number of determinations is indicated in parentheses.
98
Discussion Neither HeLa nor R2 cells contain the rat pituitary-specific factor (Catanzaro et al., 1987; Lefevre et al., 1987; Lemaigre et al., 1989) and thus no expression of hGH-N, hCS-A or rGH would be expected in these cells. This is supported by data from previous studies in which significant expression from the hGH-N, hCS-A or rGH promoters was observed in rat pituitary cells only (Cattini et al., 1986b; Cattini and Eberhardt, 1987; Lefevre et al., 1987). However, the results of this study suggest that this represents preferential expression of the hGH-N and hCS-A genes in rat pituitary cells, since these genes are active, but at a lower level, in rat non-pituitary, R2, cells after gene transfer (Fig. 3). However, no hGH-N gene expression could be detected in HeLa cells (Fig. 4). In contrast, - 109Kp.hGH-N gene was expressed in HeLa cells, which suggests that the hGH-N mRNA is stable in HeLa cells and that inactivation occurs at the transcriptional level (Fig. 4). The expression of hGH-N and hCS-A in R2 cells is the result of either random transcription initiation or promoter activation. The results are not consistent with hGH-N and hCS-A expression in R2 cells resulting solely from random initiation of transcription. Firstly, the hGH-N and hCS-A transcripts are detected, while rGHm and -234rGHm mRNAs are absent in R2 cells after gene transfer (Fig. 3). Secondly, the level of promoterless gene expression in R2 cells is an indication of the amount of random transcription initiation. Hybrid genes containing the hGH-N or hCS-A promoter sequences were more active than a promoterless gene in R2 cells (Table 1). By contrast, the hybrid rGH gene displayed a comparable level of activity with the promoterless gene (Table 1). One interpretation of these results is that rat transcription factors other than GHF-1 can activate the hGH-N and hCS-A genes. The presence of an activator other than GHF-1 for the hGH-N or hCS-A gene suggests that differential repression may play an important role in tissuespecific expression of the human growth hormone family. The information required for this repression would be contained within the 496 bp of the hGH-N or hCS-A 5’-flanking DNA used in this
study (Fig. 4; Table I). Expression of the placental members of the hGH gene family would be blocked in the pituitary through binding of a pituitary-specific repressor to unique sequences associated with the placental genes. The recent sequencing of the complete hGH gene family locus on chromosome 17 has identified a potential candidate for a placenta-specific element and factor binding domain. These are the ‘p sequences’ which are found upstream of all the hGH family members expressed in the placenta but not in pituitary-specific hGH-N (Chen et al., 1989). Expression of hGH-N in the placenta would be blocked in a similar manner by a placenta-specific repressor. This differs from the situation in the rat where rGH expression is governed by the presence of a rat pituitary-specific activator, variously called GHF-1, Pit 1, GC 1 or 2. The absence of this factor in a tissue or cell is sufficient to block expression as indicated in the results of this and other studies. Alternatively, the ability of rat pituitary and non-pituitary transcription factors to activate hGH-N and hCS-A (Fig. 3; Table 1) may simply reflect a promiscuous property of rat transcriptional activators when placed in a heterologous system, and would not occur with the equivalent human factors. At its simplest, this would imply that the human equivalent to GHF-1 is sufficiently different from the rat pituitary-specific factor, that it is able to distinguish between hGH-N and hCS-A (Nachtigal et al., 1989). The hGH-N gene would not be expressed in the placenta because no pituitary-specific factor is present (Chen et al., 1989). No repressor is required as the hGHN gene would be unable to interact with the, as yet unidentified, placenta-specific factor which acts, possibly through the ‘p sequences’. Similarly, the reciprocal situation would apply to the hCS-A gene and the other placenta-specific members of the hGH gene family in the pituitary; no placenta-specific factor would be present. This interpretation is more attractive than the first since tissue-specific expression of the hGH-N, hCS-A and rGH genes could be accomplished in the same manner, i.e., through the simple presence of pituitaryor placenta-specific activators. However, this does require that rat (GHF-1) and human pituitary-specific factors possess different
99
properties, i.e., the ability to distinguish hGH-N and hCS-A. Consequently, some caution would be indicated when extrapolating from a heterologous rat/ human to a homologous human system. Acknowledgements This work was supported by grants from the Medical Research Council of Canada (MRCC) and the Manitoba Health Research Council. B.E. Nickel is a recipient of an MRCC Fellowship and P.A. Cattini is a recipient of an MRCC Scholarship. References Catanzaro, D.F., West, B.L., Reudelhumber, T.L. and Baxter, J.D. (1987) Mol. Endocrinol. 1, 90-96. Cattini, P.A., Anderson, T.R., Baxter, J.D. and Eberhardt, N.L. (1986a) J. Biol. Chem. 261, 13367-13372. Cattini, P.A., Peritz, L.P., Anderson, T.R., Baxter, J.D. and Eberhardt, N.L. (1986b) DNA 5, 503-509. Cattini, P.A., Klassen, M.E. and Nachtigal, M.N. (1988) Mol. Cell. Endocrinol. 60, 217-224. Chen, E.Y., Liao, Y.-C., Smith, D.H., Barrera-Saldana. H.A., Gelinas, R.E. and Seeburg, P.A. (1989) Genomics 4, 479-497. Chomczynski, P. and Sac&i, N. (1987) Anal. Biochem. 162, 156-159. Eberhardt. N.L., Hirt, H., Cattini, P.A., Anderson, T., Peritz,
L., Baxter, J.D., Mellon, P., Isaacs, R., Slater, E.P. and Barta, A. (1988) in Endocrine Genes: Analytical Methods, Experimental Approaches and Selected Systems (Lau, Y.-F., ed.), pp. 149-168, Oxford University Press, New York. Klassen, M.E., Nachtigal, M.W. and Cattini, P.A. (1989) Placenta 10, 321-329. Lefevre, C., Imagawa, M., Dana, S., Grindlay, J., Bodner, M. and Karin, M. (1987) EMBO J. 66, 971-981. Lemaigre, F.P., Peers, B., Lafontaine, D.A., Marthay-Hartet, M., Rousseau, G.G., Belayew, A. and Martial, J.A. (1989) DNA 8, 1499159. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) in Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratories, Cold Spring Harbor, NY. Martial, J.A., Hallewell, R.A., Baxter, J.D. and Goodman, H.M. (1979) Science 205, 602-607. Miller, W.L. and Eberhardt, N.L. (1983) Endocr. Rev. 4, 97-130. Nachtigal, M.N., Nickel, B.E., Klassen, M.E., Zhang, W., Eberhardt, N.L. and Cattini, P.A. (1989) Nucleic Acids Res. 17, 4327-4337. Neumann, J.R., Morency, CA. and Russian, K.O. (1987) Biotechniques 5, 44-447. Nickel, B.E., Kardami, E. and Cattini, P.A. (1990) Endocrinology (in press). Peritz, N.L., Fodor, E.J.B., Silversides, D.W., Cattini, P.A., Baxter, J.D. and Eberhardt, N.L. (1988) J. Biol. Chem. 263, 5005-5007. Seeburg, P.H., Shine, J., Martial, J.A., Baxter, J.D. and Goodman, H.M. (1977) Nature 270, 486-494. West, B.L., Catanzaro, D.F., Mellon, S.H., Cattini, P.A., Baxter, J.D. and Reudelhuber, T.L. (1987) Mol. Cell. Biol. 7, 1193-1197.
MOLCEL
93
02231
Human growth hormone gene expression in rat but not human non-pituitary cells after stable gene transfer Barbara E. Nickel, Mark N. Nachtigal,
Margaret
E. Klassen and Peter A. Cattini
Department of Physiology, University of Manitoba, Winnipeg, Manitoba R3E 0 W3, Canada (Received
Key words: Growth
hormone;
Tissue-specific
23 August
expression;
1989; accepted
22 November
1989)
Gene transfer
Summary Tissue-specific expression of the rat growth hormone (rGH) gene requires binding of a pituitary-specific factor. Binding of this factor has been used to explain tissue-specific expression of the human growth hormone (hGH-N) gene in transfected rat pituitary (GC) tumour cells. Neither rat fibroblast (R2) nor human cervical carcinoma (HeLa) cells contain the rat pituitary-specific factor. Thus, no expression of hGH-N or rGH would be expected in these cells. R2 cell lines containing stably integrated hGH-N or rGH genes were generated. Expression of hGH-N but not rGH was detected. By contrast, stably transfected HeLa cells did not express the endogenous or transfected hGH-N genes. However, an hGH-N transcript was detected when hGH-N gene expression was directed by a viral promoter. This suggests that the block in expression occurs at the level of transcription and not mRNA stability. Hybrid genes containing 496 base pairs (bp) of hGH-N or 234 bp of rGH 5’-flanking DNA, including promoter sequences, fused to the bacterial gene coding for chloramphenicol acetyltransferase were used to stably transfect R2 cells. The hybrid hGH-N gene was more active than a promoterless construction in these cells. By contrast, the hybrid rGH gene was not. These data suggest that the hGH-N gene can be activated by rat transcription factors other than those found in pituitary cells.
Introduction Human pituitary growth hormone (hGH-N) and chorionic somatomamrnotropin (hCS) are two members of the human growth hormone hGH family (Eberhardt et al., 1988). The genes coding for these hormones share greater than 90% nucleotide sequence homology yet production of hGH-N is limited to the somatotrophs of the
Address for correspondence: Barbara E. Nickel, Department of Physiology, University of Manitoba, 770 Bannatyne Avenue, Winnipeg, Manitoba R3E 0W3, Canada. 0303-7207/90/$03.50
0 1990 Elsevier Scientific
Publishers
Ireland,
anterior pituitary and hCS to the syncytiotrophoblast of the placenta (Miller and Eberhardt, 1983). The mechanism by which pituitary- and placenta-specific members of the hGH gene family are differentially expressed in a tissue-specific manner is still unclear (Chen et al., 1989; Nachtigal et al., 1989). Preferential expression of the rat growth hormone (rGH) gene and the hGH-N gene is obtained in rat anterior pituitary tumour (GC) cells compared to non-pituitary cells after gene transfer (Cattini et al., 1986). This activity has been explained by the presence of a pituitary-specific factor variously called GHF-1, pit 1 GCl or 2 in these cells (Catanzaro et al., Ltd
94
1987; Lefevre et al., 1987; West et al.. 1987). The absence of this factor in non-pituitary cells has led to the conclusion that binding of GHF-1 results in tissue-specific expression of hGH-N as well as rGH. However, two placenta-specific members of the hGH gene family, hCS and placental growth hormone variant (hGH-V), are also expressed in GC cells after gene transfer (Nachtigal et al., 1989; Nickel et al., 1990). Binding of GHF-1 to the hCS-A or hCS-B genes which are both able to code for hCS has been demonstrated (Lemaigre et al., 1989; Nachtigal et al., 1989) and, in the case of hCS-A, is responsible for promoter activity in transfected GC cells (Nachtigal et al., 1989). A possible explanation to help reconcile these data is that tissue-specific expression of the hGH gene family might be accomplished through differential repression. There is evidence for repressor activity in the 5’flanking DNA of the hGH-N (Peritz et al., 1988) and hCS-A genes (Nachtigal et al., 1989). The hGH/hCS genes would be activated by factors such as GHF-1 or their non-pituitary cell equivalents, and silenced by pituitaryor placenta-specific repressors. As a consequence, members of the hGH gene family might be activated by factors other than GHF-1, if the appropriate repressors are absent. To investigate the role of repression and the possibility of additional activating factors, we examined hGH-N gene expression in non-pituitary rat and human cells after gene transfer. Our results suggest that the hGH-N gene can be activated by rat transcription factors other than GHF-1, and suggest that efficient silencing of the hGH-N gene occurs in human but not rat nonpituitary cells. These data are discussed in relation to current ideas on tissue-specific expression of the hGH gene family.
(-496hGH-Np.car). hCS-A (-496hCS-Ap.cat) or 234 bp of rGH 5’-flanking DNA ( - 234rGHp.cat) fused upstream of the bacterial gene coding for chloramphenicol acetyltransferase (CAT). All constructions were made using protocols described in Maniatis et al. (1982) unless otherwise stated. The plasmid containing both the hCS-A and neomycin resistance gene (pneohCS-A) was constructed as previously described for pneohGHN but using the hCS-A gene in place of the hGH-N gene (Cattini et al., 1986a). The truncated rGHm gene was generated by deleting 5’-flanking sequences -235 (BgfII) to - 1753 (EcoRI) from prGHm (Cattini et al., 1986a). A hybrid gene was made containing the herpes-simplex thymidine kinase (TIS) promoter (p.) fused upstream of the hGH-N structural gene sequences. The plasmid phGH-N (Cattini et al., 1986a) was cut with BamHI to remove the 5’-flanking and promoter This fragment was replaced with a sequences. BamHI/Bg/II promoter fragment of TK containing nucleotides - 109 to + 53 (Cattini et al., 1986b) to generate - 109TKp.hGH-N.
Cell culture and gene transfer Human cervical carcinoma (HeLa), rat fibroblast (R2) and anterior pituitary tumour (GC) cells were grown in monolayer culture in Dulbecco’s modified Eagle’s medium supplemented with 2.5% fetal calf serum and 2.5% calf serum. For stable gene transfer, 2-3 X lo6 cells were plated per 100 mm culture dish and transfections were done by calcium phosphate-DNA precipitation as described previously (Cattini et al., 1988). Resistant clones were selected with 0.5 mg/ml (active) G-418 (Sigma) and a number greater than 50 was designated as a pool.
DNA isolation and blotting Materials and methods
Materials Many of the gene or hybrid gene fragments used in this study have been described previously (Cattini et al., 1986a, b; Nachtigal et al., 1989). These include hGH-N, hCS-A, modified rGH (rGHm), the gene coding for neomycin resistance (neodX), neohGH-N, neorGHm, hybrid genes containing 496 base pairs (bp) of hGH-N
Medium was aspirated from the culture dish and the cells washed with 5 ml of phosphatebuffered saline (PBS). The cells were scraped and transferred to a 15 ml tube in a further 5 ml of PBS. The cells were pelleted by centrifugation, and resuspended in 1 ml of PBS before transfer to a 1.5 ml microfuge tube. The suspension was spun for 10 s to pellet the cells and the PBS was aspirated. The cells were lysed by adding 1 ml of 10 mM Ttis-HCl pH 7.6,150 mM sodium chloride
95
(NaCl), 2 mM magnesium chloride (MgCl,) and 0.5% NP40 with incubation on ice for 5 min. The nuclei were spun for 20 s in a microfuge and the supernatant was discarded. The nuclear pellet was resuspended in 0.5 ml of 15 mM Tris-HCI pH 7.5, 160 mM NaCI, 3 mM MgCl,, 10 mM EDTA and 2.2% Na-sarcosyl and then 25 ~1 of 10 mg/ml pronase was added and the mixture incubated overnight at 37” C. The digest was then treated with 10 mg/ml ribonuclease for 3 h followed by 25 ~1 of 10 mg/ml pronase with overnight incubation at 37’C. The protein was removed from the digest by repeated phenol : chloroform steps until the interface was clear (at least three) and the DNA pr~pitated with isopropanol (Maniatis et al., 1982). The DNA was resuspended in 10 mM Tris-HCI pH 7.5, 1 mM EDTA. DNA (5 pg) was digested with restriction endonucleases and then fractionated in 1% agarose gels, transferred to nitrocellulose and assessed with radiolabelled probes according to established protocols (Maniatis et al., 1982). RNA isolation and blotting Total cytoplasmic RNA was isolated by a modification of the procedure of Chomczynski and Sac&i (1987) as described elsewhere (Klassen et al., 1989). Total cytoplasmic RNA was denatured with formaldehyde and fractionated by electrophoresis through a 1.5% agarose gel (Maniatis et al., 1982). The RNA was blotted onto nitrocellulose and probed with 32P-radiolabelled hGH cDNA (Martial et al., 1979), rGH cDNA (Seeburg et al., 1977) or B-actin (kindly provided by Dr. S. deToledo) at 42°C in the presence of 50% formamide. Blots were wahsed 3 times for 15 min in 0.1 x SSC (20 x SSC; 3 M NaCl, 0.3 M sodium citrate) with 0.1% sodium dodecyl sulphate at 65OC and then assessed by autora~o~aphy. To remove the previous probe, the nitrocellulose was washed 3 times with boiling water before subsequent prehybridization and hybridization steps. CA T assays CAT assays were carried out as previously described (Nachtigal et al., 1989) and activity was measured using two-phase fluor diffusion assay (Neumann et al., 1987). Quantitative values for CAT activity were determined by regression anal-
ysis to give counts per minute, per minute, per milligram (cpm/minfmg) of cell protein extract. Results
Expression of hGH-N, KS-A, rGHm and -234rGHm in stably transfected non-pituitary R2 cells Modified rGH genes, rGHm, neorGHm and - 234rGHm, as well as hGH-N, neohGH-N, hCSA and neohCS-A were stably introduced into R2 cells to generate the cell lines R2[rGHm], RZ[neorGHm], R2[ - 234rGHm], R2[hGH-N], R2[neohGH-N], R2[hCS-A] and R2[neohCS-A]. The modification in the rGH genes was made by introducing a BamHI decanucleotide linker into the 5 ‘-untranslated region. This modification allows us to distinguish between the transfected and endogenous rGH gene. We have shown previously that rGHm is expressed along with the endogenous rGH gene in GC cells after gene transfer (Cattini et al., 1986a). Analysis of genomic DNA from the cell lines generated was done to determine that successful integration of the transfected genes had occurred. Genomic DNA from each cell line was isolated and cut with restriction endonucleases which identify the transfected gene; EcoRI or PuuII for hGH-N and hCS-A genes, and Hind111 compared to Hind111 with BarnHI for the endogenous and transfected rGH genes. The DNA was resolved by electrophoresis, transferred to nitrocellulose, probed with radiolabelled hGH cDNA for the human genes and rGH cDNA for the rGH genes, and assessed by autoradiography (Figs. 1 and 2). The integration of hGH-N or hCS-A in the R2 genome is indicated by hybridization of DNA to the hGH cDNA probe compared to untransfect~ R2 DNA, and the presence of approp~ately sized fragments (Fig. 1). The integration of the modified rGH genes is indicated by the presence of additional bands in the transfected compared to the untransfected cell DNA that hybridizes to rGH cDNA. Unlike the endogenous rGH gene DNA which contains no BamHI site within its structural sequences, the bands corresponding to rGHm or - 234rGHm are cleaved by BamHI (Fig, 2). Total cytoplasmic RNA isolated from R2 cell lines containing hGH-N, hCS-A, rGHm and
96
-234rGH, as well as untransfected R2 cells was assessed by RNA blotting. RNA was separated by transferred to nitrocellulose, electrophoresis, probed sequentially with radiolabelled hGH cDNA, rGH cDNA and P-actin, and each was visualized by autoradiography (Fig. 3). RNA from untransfected GC cells and GC cells transfected with the hGH-N gene was also analyzed as controls (Fig. 3). Both hGH-N and hCS-A gene expression was observed in transfected R2 cells. By contrast, neither rGHm, - 234rGH or endogenous rGH mRNA was detected (Fig. 3). Stability of hGH-N transcripts in stably transfected HeLu cells Human cervical carcinoma (HeLa) cells were stably transfected with the hGH-N gene and four separate pools (HeLa[hGH-N]l-4) generated. No
abcdefghij Fig. 1. Detection of hGH-N or hCS-A genes in R2 cells by DNA blotting. DNA (5 pg) isolated from untransfected R2 cells (a,b) or R2 cells stably transfected with hGH-N (cd), hCS-A (e,f), neohGH-N (gh) and neohCS-A genes (ij) was cut with EcoRI (a,c,e,g,i) or PuuII (b,d,f,hj). The DNA was fractionated in a 1% agarose gel, transferred to nitrocellulose, probed with radiolabelled hGH-N cDNA and assessed by autoradiography. No hybridization to untransfected R2 DNA is observed (a,b), and bands corresponding to 2.6 kb hGH-N (c,g) and 2.9 kb hCS-A (e,i) EcoRI fragments, respectively, are seen.
abcdef Fig. 2. Detection of transfected and endogenous rGH genes in R2 cells by DNA blotting. DNA (5 pg) from untransfected R2 cells (a,b) and R2 cells transfected with the rGHm (cd) and -234rGHm genes (e,f) was cut with Hind111 (a,c,e) and HindIII/BumHI (b,d,f). The DNA was resolved in a 1% agarose gel, transferred to nitrocellulose and probed with a radiolabelled rGH cDNA and assessed by autoradiography. Additional bands are seen in the DNA from transfected cells which are cut by BarnHI; BamHI is the site used to make the modification in the rGH gene structural sequences.
endogenous or transfected hGH-N mRNA was detected in 0.1 mg of total cytoplasmic RNA from any of these lines by RNA blotting analysis, even after prolonged autoradiographic exposure (data not shown). This could be due to an undetectable level of hGH-N transcription or instability of hGH-N transcripts in HeLa cells. To determine which of these possibilities are most probable, a hybrid gene( - 109TKp.hGH-N) was constructed in which the hGH-N gene 5’-flanking DNA, including the endogenous promoter, was replaced with the herpes-simplex thymidine kinase promoter( - 109TKp). The - 109TKp.hGH-N gene was stably introduced into HeLa cells and expression assessed by RNA blotting analysis. RNA from HeLa[hGH-N]l and untransfected HeLa cells were included as controls (Fig. 4). A - 109TKp.hGH-N transcript was detected in HeLa cells suggesting that hGH-N mRNA would be stable in these cells if produced. These data suggest that the absence of endogenous or transfected hGH-N mRNA in HeLa cells reflects an effect at the level of transcription and not mRNA stability.
97
28s
18s
a
bc
Fig. 4. Expression of hGH-N in HeLa cells by RNA blotting. Total cytoplasmic RNA was isolated from HeLa cells transfected with (a) - 109TKp.hGH-N or (b) hGH-N (pool 1) and (c) untransfected HeLa cells. RNA (0.1 mg) was resolved in a formaldehyde-agarose gel, transferred to nitrocellulose and probed with radiolabelled hGH-N cDNA, and assessed by autoradiography.
a
bcdefghijk
Fig. 3. Expression of transfected hGH-N, hCS-B and rGH in R2 cells detected by RNA blotting. Total cytoplasmic RNA was isolated from (a) GC[hGH-N], (b) GC, (c,f,k) R2, (d) RZ[hGH-N], (e) RZ[hCS-A], (g) RZ[neohGH-N], (h) IU[neohCS-A], (i) IU[rGHm] and (i) R2[ -234rGHml cells. RNA (20 pg) was fractionated in a formaldehyde-agarose gel, transferred to nitrocellulose and probed with radiolabelled hGH-N, rGH and B-actin cDNAs as indicated, and assessed by autoradiography.
The role of proximal hGH-N and hCS-A S’-jlanking sequences on gene repression Hybrid cat genes were made containing 496 bp of hGH-N, hCS-A or 234 bp of rGH 5’-flanking sequences fused upstream of the bacterial gene (cat) coding for chloramphenicol acetyltransferase (CAT). These genes are expressed efficiently in GC cells after gene transfer (Cattini et al., 1986; Cattini and Eberhardt, 1987). The same hybrid cat genes were used to stably transfect B2 cells, and expression was assessed by measuring CAT activity in a two-phase fluor diffusion assay. A cell line containing a promoterless (-p) cat gene was also
made (B2[ -p.cat] cells) and assayed to give an indication of the level for random initiation of transcription (Table 1). Hybrid hGH-N and hCS-A genes were more active than the -p.cat gene in R2 cells. By comparison, the -234rGHp.cat gene was not expressed. TABLE
1
EXPRESSION OF HYBRID cut GENES DIRECTED hGH-N, hCS-A OR rGH 5’-FLANKING SEQUENCES STABLY TRANSFECTED R2 CELLS Cell line
Gene
cpm/min/mg
R2 R2[ R2[ R2[ R2[
- p.caf - 234rGHp. cur -496hGH-Np.cur -496hCSAp.cat
77.1+ 1os.9+ 94.5 + 283.4+ 347.2 +
-p.car] - 234rGHp. car] - 496hGH-Np.car] -496hCS-Ap.cat]
3.6 4.4 2.4 2.1 1.3
BY IN
(4) (3) (3) (4) (3)
CAT activity was assessed in untransfected R2 cells and R2 cells stably transfected with a promoterless cur gene ( - p. cut), or hybrid car genes containing 234 or 496 bp of ffiH or hGH-N and hCS-A 5’-flanking DNA, respectively. Values for expression are given as cpm/min/mg cell extract plus or minus standard error from the mean. The number of determinations is indicated in parentheses.
98
Discussion Neither HeLa nor R2 cells contain the rat pituitary-specific factor (Catanzaro et al., 1987; Lefevre et al., 1987; Lemaigre et al., 1989) and thus no expression of hGH-N, hCS-A or rGH would be expected in these cells. This is supported by data from previous studies in which significant expression from the hGH-N, hCS-A or rGH promoters was observed in rat pituitary cells only (Cattini et al., 1986b; Cattini and Eberhardt, 1987; Lefevre et al., 1987). However, the results of this study suggest that this represents preferential expression of the hGH-N and hCS-A genes in rat pituitary cells, since these genes are active, but at a lower level, in rat non-pituitary, R2, cells after gene transfer (Fig. 3). However, no hGH-N gene expression could be detected in HeLa cells (Fig. 4). In contrast, - 109Kp.hGH-N gene was expressed in HeLa cells, which suggests that the hGH-N mRNA is stable in HeLa cells and that inactivation occurs at the transcriptional level (Fig. 4). The expression of hGH-N and hCS-A in R2 cells is the result of either random transcription initiation or promoter activation. The results are not consistent with hGH-N and hCS-A expression in R2 cells resulting solely from random initiation of transcription. Firstly, the hGH-N and hCS-A transcripts are detected, while rGHm and -234rGHm mRNAs are absent in R2 cells after gene transfer (Fig. 3). Secondly, the level of promoterless gene expression in R2 cells is an indication of the amount of random transcription initiation. Hybrid genes containing the hGH-N or hCS-A promoter sequences were more active than a promoterless gene in R2 cells (Table 1). By contrast, the hybrid rGH gene displayed a comparable level of activity with the promoterless gene (Table 1). One interpretation of these results is that rat transcription factors other than GHF-1 can activate the hGH-N and hCS-A genes. The presence of an activator other than GHF-1 for the hGH-N or hCS-A gene suggests that differential repression may play an important role in tissuespecific expression of the human growth hormone family. The information required for this repression would be contained within the 496 bp of the hGH-N or hCS-A 5’-flanking DNA used in this
study (Fig. 4; Table I). Expression of the placental members of the hGH gene family would be blocked in the pituitary through binding of a pituitary-specific repressor to unique sequences associated with the placental genes. The recent sequencing of the complete hGH gene family locus on chromosome 17 has identified a potential candidate for a placenta-specific element and factor binding domain. These are the ‘p sequences’ which are found upstream of all the hGH family members expressed in the placenta but not in pituitary-specific hGH-N (Chen et al., 1989). Expression of hGH-N in the placenta would be blocked in a similar manner by a placenta-specific repressor. This differs from the situation in the rat where rGH expression is governed by the presence of a rat pituitary-specific activator, variously called GHF-1, Pit 1, GC 1 or 2. The absence of this factor in a tissue or cell is sufficient to block expression as indicated in the results of this and other studies. Alternatively, the ability of rat pituitary and non-pituitary transcription factors to activate hGH-N and hCS-A (Fig. 3; Table 1) may simply reflect a promiscuous property of rat transcriptional activators when placed in a heterologous system, and would not occur with the equivalent human factors. At its simplest, this would imply that the human equivalent to GHF-1 is sufficiently different from the rat pituitary-specific factor, that it is able to distinguish between hGH-N and hCS-A (Nachtigal et al., 1989). The hGH-N gene would not be expressed in the placenta because no pituitary-specific factor is present (Chen et al., 1989). No repressor is required as the hGHN gene would be unable to interact with the, as yet unidentified, placenta-specific factor which acts, possibly through the ‘p sequences’. Similarly, the reciprocal situation would apply to the hCS-A gene and the other placenta-specific members of the hGH gene family in the pituitary; no placenta-specific factor would be present. This interpretation is more attractive than the first since tissue-specific expression of the hGH-N, hCS-A and rGH genes could be accomplished in the same manner, i.e., through the simple presence of pituitaryor placenta-specific activators. However, this does require that rat (GHF-1) and human pituitary-specific factors possess different
99
properties, i.e., the ability to distinguish hGH-N and hCS-A. Consequently, some caution would be indicated when extrapolating from a heterologous rat/ human to a homologous human system. Acknowledgements This work was supported by grants from the Medical Research Council of Canada (MRCC) and the Manitoba Health Research Council. B.E. Nickel is a recipient of an MRCC Fellowship and P.A. Cattini is a recipient of an MRCC Scholarship. References Catanzaro, D.F., West, B.L., Reudelhumber, T.L. and Baxter, J.D. (1987) Mol. Endocrinol. 1, 90-96. Cattini, P.A., Anderson, T.R., Baxter, J.D. and Eberhardt, N.L. (1986a) J. Biol. Chem. 261, 13367-13372. Cattini, P.A., Peritz, L.P., Anderson, T.R., Baxter, J.D. and Eberhardt, N.L. (1986b) DNA 5, 503-509. Cattini, P.A., Klassen, M.E. and Nachtigal, M.N. (1988) Mol. Cell. Endocrinol. 60, 217-224. Chen, E.Y., Liao, Y.-C., Smith, D.H., Barrera-Saldana. H.A., Gelinas, R.E. and Seeburg, P.A. (1989) Genomics 4, 479-497. Chomczynski, P. and Sac&i, N. (1987) Anal. Biochem. 162, 156-159. Eberhardt. N.L., Hirt, H., Cattini, P.A., Anderson, T., Peritz,
L., Baxter, J.D., Mellon, P., Isaacs, R., Slater, E.P. and Barta, A. (1988) in Endocrine Genes: Analytical Methods, Experimental Approaches and Selected Systems (Lau, Y.-F., ed.), pp. 149-168, Oxford University Press, New York. Klassen, M.E., Nachtigal, M.W. and Cattini, P.A. (1989) Placenta 10, 321-329. Lefevre, C., Imagawa, M., Dana, S., Grindlay, J., Bodner, M. and Karin, M. (1987) EMBO J. 66, 971-981. Lemaigre, F.P., Peers, B., Lafontaine, D.A., Marthay-Hartet, M., Rousseau, G.G., Belayew, A. and Martial, J.A. (1989) DNA 8, 1499159. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) in Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratories, Cold Spring Harbor, NY. Martial, J.A., Hallewell, R.A., Baxter, J.D. and Goodman, H.M. (1979) Science 205, 602-607. Miller, W.L. and Eberhardt, N.L. (1983) Endocr. Rev. 4, 97-130. Nachtigal, M.N., Nickel, B.E., Klassen, M.E., Zhang, W., Eberhardt, N.L. and Cattini, P.A. (1989) Nucleic Acids Res. 17, 4327-4337. Neumann, J.R., Morency, CA. and Russian, K.O. (1987) Biotechniques 5, 44-447. Nickel, B.E., Kardami, E. and Cattini, P.A. (1990) Endocrinology (in press). Peritz, N.L., Fodor, E.J.B., Silversides, D.W., Cattini, P.A., Baxter, J.D. and Eberhardt, N.L. (1988) J. Biol. Chem. 263, 5005-5007. Seeburg, P.H., Shine, J., Martial, J.A., Baxter, J.D. and Goodman, H.M. (1977) Nature 270, 486-494. West, B.L., Catanzaro, D.F., Mellon, S.H., Cattini, P.A., Baxter, J.D. and Reudelhuber, T.L. (1987) Mol. Cell. Biol. 7, 1193-1197.
