Interactions between Rat Prolactin Gene Promoter and Enhancer Regions in Mammosomatotrope and Lactotrope Cell Lines
Graham J. Jones and Daniel F. Catanzaro School of Biological Sciences University of Sydney Sydney 2006, Australia
Although both promoter and enhancer sequences of the PRL gene 5'-flanking DNA are required for cellspecific, high level expression in transgenic animals, reports of the relative contributions of these elements determined in transient transfection experiments have varied. In this study we examined the transcriptional activities of proximal promoter (-422/+33) and distal enhancer (-1956/-1530) sequences of the rat (r) PRL gene by transient transfection of hybrid genes containing these sequences into two rat pituitary cell lines, GC and 235-1. These cell lines exhibit characteristics either of mammosomatotropes, which express both PRL and the evolutionarily related GH gene (GC), or lactotropes, which express only PRL (235-1). As lactotropes are thought to differentiate from a mammosomatotrope precursor cell, comparisons between these cell lines provide the opportunity to examine the mechanisms that activate PRL and GH genes during development. We show that the relative contributions of promoter and enhancer elements differ between GC and 2351 cells. Although maximal enhancer-driven activity was similar between these cell lines, promoter sequences were more active in GC (5-10% maximal) than 235-1 cells (1-2% maximal). However, no apparent differences in factor binding to the rPRL promoter region could be correlated with differences in activity, suggesting that differential factor modification, rather than different factors, is involved. As the rGH promoter exhibited a similar pattern of activity in these cell lines, these observations suggest that promoter as well as enhancer elements contribute to the cell specificity of PRL and GH gene expression. The identification of a strain of 235-1 cells that more closely resembles GC cells in its utilization of rPRL and rGH promoter and enhancer elements suggests that the lactotrope phenotype may revert to that of the mammosomatotrope precursor. The effects of such differences between strains of cell lines and other factors affecting the
apparent activity of rPRL promoter and enhancer elements are discussed. (Molecular Endocrinology 5: 1836-1844, 1991)
INTRODUCTION
The expression of PRL and GH genes by distinct cell types of the anterior pituitary (lactotropes and somatotropes, respectively) provides an excellent model system to study the mechanisms that direct cell-specific expression. Lactotropes and somatotropes are thought to arise from a common mammosomatotrope precursor cell that expresses both PRL and GH genes (1, 2). Coexpression of PRL and GH by mammosomatotropes (3) suggests that a common factor(s) may be involved in their transcriptional control. Accordingly, a pituitaryspecific factor, Pit-1, has been cloned (4) and shown to stimulate the expression of both PRL and GH (5). Although an identical factor cloned by a different laboratory (GHF-1) (6) appears to stimulate only GH expression, at least one other laboratory (7) has independently demonstrated that Pit-1 /GHF-1 activates the expression of both PRL and GH genes. Pit-1/GHF-1 has been shown to bind four sites in the rat (r) PRL proximal 5'-flanking DNA region (-422/+33) and another four sites in the distal region (-1956/ -1530) (8). While either region is sufficient to direct the expression of reporter genes to pituitary cells of transgenic animals, both regions are required for high level expression confined to lactotropes (9). Thus, synergistic interactions between proximal and distal elements have been implicated in determining cell-specific expression of the rPRL gene at physiological levels (9). In the rGH gene, Pit-1/GHF-1 binds two sites within a 180-basepair (bp) region of the 5'-flanking DNA sequence, which is sufficient for somatotrope-specific expression (10). Although transfection experiments using rat pituitary cell lines have generally supported these findings (8,11), a number of studies have found little if any activation of PRL expression by distal sequences (12-16). However, several more recent studies have
0888-8809/91 /1836-1844$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society
1836
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 03:07 For personal use only. No other uses without permission. . All rights reserved.
1837
PRL Promoter and Enhancer Interactions
demonstrated enhancer activity associated with the distal sequences of rat (17), bovine (18), and human (19) PRL genes. Of the many studies that have examined PRL and GH expression by transfection of hybrid genes, most have used one of the GH series of cell lines (20), GH3) GH4, or GC. These closely related cell lines are derived from a common rat pituitary tumor and express their endogenous PRL and GH genes to varying degrees (20). As such, these cells appear to be more closely related to the precursor mammosomatotrope that expresses both genes. Generally, members of this family of cell lines express transfected PRL and GH genes with high efficiency. Although GC cells express no detectable rPRL, their lineage from GH3 cells and their ability to utilize PRL promoter and enhancer sequences distinguish them as mammosomatotropes. Another rat cell line, 235-1, isolated from a different pituitary tumor, expresses PRL at high levels, but not GH (21), and thus more closely resembles mature lactotropes. Transfection experiments with 235-1 cells have shown rGH (3, 5,11) and rPRL proximal sequences (11) to be relatively inactive and dependent on distal sequences for activity, although some activity has been observed with rPRL proximal sequences (5, 7). The present studies were carried out to examine interactions between promoter and enhancer regions in the rPRL gene and their cell specificity. Hybrid genes containing serial deletions of the rPRL proximal promoter region in either the presence or absence of the distal enhancer were transiently transfected into mammosomatotrope (GC) and lactotrope (235-1) cells. In both cell lines the enhancer stimulated expression from proximal promoter sequences to similar levels. However, promoter sequences were considerably less active in 235-1 than in GC cells. In GC cells, the -296/ -173 promoter region was required for maximal promoter activity and maximal stimulation by the enhancer. Close parallels in the activity of rGH promoter sequences between GC and 235-1 cell lines suggest that differentiation of mammosomatotropes to lactotropes involves a repression mechanism that acts on both rGH and rPRL promoters, but is overridden by the PRL enhancer. However, our inability to detect differences between GC and 235-1 cells in factor binding to the rPRL promoter suggests that differential factor modification may be involved. Direct comparisons between two strains of 235-1 cells revealed that one strain more closely resembles GC cells in its utilization of PRL and GH promoter and enhancer elements. This indicates that the lactotrope phenotype may revert to that of its mammosomatotrope precursor and provides an explanation for differences in the activities of rPRL promoter and enhancer elements reported in the literature. Enhancer activity was also shown to be sensitive to the amount of plasmid transfected, suggesting another important variable which may account for reported differences in promoter and enhancer activities.
RESULTS
To examine the contribution of sequences in the rPRL 5'-flanking DNA to its promoter activity, a series of endpoint deletions from -1956 to - 7 9 was linked to a bacterial chloramphenicol acetyl transferase (CAT) gene. Another set of constructions was made which joined the distal sequence (-1956 to -1530) in the forward orientation to the 5'-end of each proximal promoter construct. The sequences used in these constructions and the locations of factor-binding sites are shown in Fig. 1. These constructions were transfected by electroporation into GC and two clones of 235-1 cells. In each experiment, activities were normalized to the activity of the -1956 construct (set at unity) after correction for variations in transfection efficiency (
Graham J. Jones and Daniel F. Catanzaro School of Biological Sciences University of Sydney Sydney 2006, Australia
Although both promoter and enhancer sequences of the PRL gene 5'-flanking DNA are required for cellspecific, high level expression in transgenic animals, reports of the relative contributions of these elements determined in transient transfection experiments have varied. In this study we examined the transcriptional activities of proximal promoter (-422/+33) and distal enhancer (-1956/-1530) sequences of the rat (r) PRL gene by transient transfection of hybrid genes containing these sequences into two rat pituitary cell lines, GC and 235-1. These cell lines exhibit characteristics either of mammosomatotropes, which express both PRL and the evolutionarily related GH gene (GC), or lactotropes, which express only PRL (235-1). As lactotropes are thought to differentiate from a mammosomatotrope precursor cell, comparisons between these cell lines provide the opportunity to examine the mechanisms that activate PRL and GH genes during development. We show that the relative contributions of promoter and enhancer elements differ between GC and 2351 cells. Although maximal enhancer-driven activity was similar between these cell lines, promoter sequences were more active in GC (5-10% maximal) than 235-1 cells (1-2% maximal). However, no apparent differences in factor binding to the rPRL promoter region could be correlated with differences in activity, suggesting that differential factor modification, rather than different factors, is involved. As the rGH promoter exhibited a similar pattern of activity in these cell lines, these observations suggest that promoter as well as enhancer elements contribute to the cell specificity of PRL and GH gene expression. The identification of a strain of 235-1 cells that more closely resembles GC cells in its utilization of rPRL and rGH promoter and enhancer elements suggests that the lactotrope phenotype may revert to that of the mammosomatotrope precursor. The effects of such differences between strains of cell lines and other factors affecting the
apparent activity of rPRL promoter and enhancer elements are discussed. (Molecular Endocrinology 5: 1836-1844, 1991)
INTRODUCTION
The expression of PRL and GH genes by distinct cell types of the anterior pituitary (lactotropes and somatotropes, respectively) provides an excellent model system to study the mechanisms that direct cell-specific expression. Lactotropes and somatotropes are thought to arise from a common mammosomatotrope precursor cell that expresses both PRL and GH genes (1, 2). Coexpression of PRL and GH by mammosomatotropes (3) suggests that a common factor(s) may be involved in their transcriptional control. Accordingly, a pituitaryspecific factor, Pit-1, has been cloned (4) and shown to stimulate the expression of both PRL and GH (5). Although an identical factor cloned by a different laboratory (GHF-1) (6) appears to stimulate only GH expression, at least one other laboratory (7) has independently demonstrated that Pit-1 /GHF-1 activates the expression of both PRL and GH genes. Pit-1/GHF-1 has been shown to bind four sites in the rat (r) PRL proximal 5'-flanking DNA region (-422/+33) and another four sites in the distal region (-1956/ -1530) (8). While either region is sufficient to direct the expression of reporter genes to pituitary cells of transgenic animals, both regions are required for high level expression confined to lactotropes (9). Thus, synergistic interactions between proximal and distal elements have been implicated in determining cell-specific expression of the rPRL gene at physiological levels (9). In the rGH gene, Pit-1/GHF-1 binds two sites within a 180-basepair (bp) region of the 5'-flanking DNA sequence, which is sufficient for somatotrope-specific expression (10). Although transfection experiments using rat pituitary cell lines have generally supported these findings (8,11), a number of studies have found little if any activation of PRL expression by distal sequences (12-16). However, several more recent studies have
0888-8809/91 /1836-1844$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society
1836
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 03:07 For personal use only. No other uses without permission. . All rights reserved.
1837
PRL Promoter and Enhancer Interactions
demonstrated enhancer activity associated with the distal sequences of rat (17), bovine (18), and human (19) PRL genes. Of the many studies that have examined PRL and GH expression by transfection of hybrid genes, most have used one of the GH series of cell lines (20), GH3) GH4, or GC. These closely related cell lines are derived from a common rat pituitary tumor and express their endogenous PRL and GH genes to varying degrees (20). As such, these cells appear to be more closely related to the precursor mammosomatotrope that expresses both genes. Generally, members of this family of cell lines express transfected PRL and GH genes with high efficiency. Although GC cells express no detectable rPRL, their lineage from GH3 cells and their ability to utilize PRL promoter and enhancer sequences distinguish them as mammosomatotropes. Another rat cell line, 235-1, isolated from a different pituitary tumor, expresses PRL at high levels, but not GH (21), and thus more closely resembles mature lactotropes. Transfection experiments with 235-1 cells have shown rGH (3, 5,11) and rPRL proximal sequences (11) to be relatively inactive and dependent on distal sequences for activity, although some activity has been observed with rPRL proximal sequences (5, 7). The present studies were carried out to examine interactions between promoter and enhancer regions in the rPRL gene and their cell specificity. Hybrid genes containing serial deletions of the rPRL proximal promoter region in either the presence or absence of the distal enhancer were transiently transfected into mammosomatotrope (GC) and lactotrope (235-1) cells. In both cell lines the enhancer stimulated expression from proximal promoter sequences to similar levels. However, promoter sequences were considerably less active in 235-1 than in GC cells. In GC cells, the -296/ -173 promoter region was required for maximal promoter activity and maximal stimulation by the enhancer. Close parallels in the activity of rGH promoter sequences between GC and 235-1 cell lines suggest that differentiation of mammosomatotropes to lactotropes involves a repression mechanism that acts on both rGH and rPRL promoters, but is overridden by the PRL enhancer. However, our inability to detect differences between GC and 235-1 cells in factor binding to the rPRL promoter suggests that differential factor modification may be involved. Direct comparisons between two strains of 235-1 cells revealed that one strain more closely resembles GC cells in its utilization of PRL and GH promoter and enhancer elements. This indicates that the lactotrope phenotype may revert to that of its mammosomatotrope precursor and provides an explanation for differences in the activities of rPRL promoter and enhancer elements reported in the literature. Enhancer activity was also shown to be sensitive to the amount of plasmid transfected, suggesting another important variable which may account for reported differences in promoter and enhancer activities.
RESULTS
To examine the contribution of sequences in the rPRL 5'-flanking DNA to its promoter activity, a series of endpoint deletions from -1956 to - 7 9 was linked to a bacterial chloramphenicol acetyl transferase (CAT) gene. Another set of constructions was made which joined the distal sequence (-1956 to -1530) in the forward orientation to the 5'-end of each proximal promoter construct. The sequences used in these constructions and the locations of factor-binding sites are shown in Fig. 1. These constructions were transfected by electroporation into GC and two clones of 235-1 cells. In each experiment, activities were normalized to the activity of the -1956 construct (set at unity) after correction for variations in transfection efficiency (
