The Tissue-Specific Mammalian Transcription Factor, Pit-1, Activates Transcription in Saccharomyces cerevisiae
Yi Ding*, Wei Lu, Mark S. Roberson, W. Scott Moye-Rowley, and Richard A. Maurer Department of Physiology and Biophysics University of Iowa Iowa City, Iowa 52242
Pit-1 is a tissue-specific transcription factor which binds to specific DNA sequences within 5' flanking regions of the PRL and GH genes and activates the transcription of these genes. Previous studies have shown that expression of Pit-1 is necessary to activate transcription from the PRL or GH promoters in heterologous mammalian cells. In the present study the ability of Pit-1 expression vectors to activate expression of reporter genes in Saccharomyces cerevisiae was examined. The test system used Pit1 expression vectors and an indicator plasmid containing multiple copies of a Piti-binding site as a replacement for the upstream activator sequence of the CYC1 promoter. Significant activation of indicator plasmid expression was detected only in the presence of functional Pit-1 expression vectors. In both mammalian and yeast cells, amino-terminal deletions of the Pit-1 coding sequence produced similar and gradual loss of transcriptional activation. This finding indicates that similar or identical regions of Pit-1 are required for transcriptional activation in mammalian and yeast cells. Although synthetic DNA elements containing multiple copies of a single Pit1-binding site were sufficient to permit Pit-1-mediated transcriptional activation in both yeast and mammalian cells, DNA fragments representing the proximal region or distal enhancer region of the PRL gene were transcriptionally active only in mammalian cells. These studies establish the ability of Pit1 to stimulate transcription in the absence of other tissue-specific factors and provide a system for further genetic studies of Pit-1 structure/function relationships as well as evaluation of target sequences necessary for Pit-1 action. (Molecular Endocrinology 5: 1239-1245, 1991)
INTRODUCTION The pituitary-specific transcription factor Pit-1 is involved in the activation of specific genes within the 0888-8809/91/1239-1245$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society
anterior pituitary. Although there has been some controversy concerning the target genes that are activated by Pit-1 (1, 2), several studies suggest that Pit-1 can frans-activate both the PRL and GH genes (3-6). Pit-1 is a member of a family of transcription factors, designated POU factors, that contain two conserved domains (7). One of the conserved domains is similar to homeobox domains, found in a variety of developmentally important transcription factors (8, 9), while the other domain is specific to the POU factors. Both the POUspecific and the homeodomain are important for sequence-specific DNA binding of the POU factors (1012). The regions of Pit-1 that are required for transcriptional activation have also been explored. The aminoterminal portion of Pit-1 appears to contain several regions that are important for transcriptional activation of the GH and PRL genes (10,13). One experimental approach that has been important for demonstrating the role of Pit1 in transcriptional activation has involved transfection of heterologous mammalian cells with Pit-1 expression vectors (3-6). These studies have shown that Pit-1 is sufficient to activate the PRL and GH promoters when general transcription factors are provided by a heterologous cell. To examine the ability of Pit-1 to function in a nonmammalian cellular environment and to develop a system that permits genetic dissection of Pit-1 structure/function relationships, we have explored the ability of Pit-1 to activate transcription in yeast. The ability of several mammalian transcription factors to function in yeast (14-20) indicates the high degree of functional conservation of the transcriptional apparatus among eukaryotes. Even factors that are spatially restricted, such as the fushi tarazu homeobox-containing protein from Drosophila (21) have been demonstrated to function as transcriptional activators in yeast. While several transcription factors from higher eukaryotes have been shown to function in yeast, it is conceivable that at least some tissue-specific factors may require additional transcription factors that are either not present or are not functionally conserved in yeast. For instance, transcription factors that function as heterodimers containing a
1239
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 05:45 For personal use only. No other uses without permission. . All rights reserved.
Vol 5 No. 9
MOL ENDO-1991 1240
tissue-specific and a nontissue-specific subunit might function in heterologous mammalian cells, but not in yeast. Furthermore, POU domain proteins have not been found in nonmetazoan organisms. Thus, even though Pit-1 can activate transcription in heterologous mammalian cells, it is conceivable that Pit-1 might not function in yeast. In the present studies we demonstrate that Pit-1 can frans-activate gene expression in Saccharomyces cerevisiae. While several aspects of Pit-1dependent transcriptional activation were found to be similar between mammalian cells and Saccharomyces cerevisiae, a difference in DNA sequences that can support activation in the two systems was observed.
Pit-1 Binding Sites
a.
1X1P
EB 4x1 P CYC1 Promoter (minus UAS)
Pit-1 Coding Sequence
pSEYC68GAL-Pit-1 Effector Plasmid
pCS10-1x1Por4x1P Indicator Plasmid
RESULTS Pit-1 Expression in Saccharomyces cerevisiae and Activation of Gene Transcription To determine whether Pit-1 can function in yeast, three different Pit-1 expression plasmids under the control of heterologous promoters were constructed. A regulated expression vector was prepared by inserting the Pit-1 cDNA down-stream of the GAL1 promoter. Pit-1 expression using this vector should be dependent on the use of galactose as the carbon source. Constitutive expression vectors were also prepared using either the alcohol dehydrogenase promoter or the DED1 promoter carried on pADH2-Pit-1 and YCp88-Pit-1, respectively. The effects of Pit-1 expression were assessed by analysis of the expression of appropriate indicator plasmids. The indicator plasmids contained varying numbers of Pit-1-binding sites inserted up-stream of the yeast CYC1 promoter linked to lacZ. The CYC1 upstream activator sequence has been deleted from this construct, so that /3-galactosidase expression is dependent on binding of transcriptional activators to the inserted Pit-1-binding sites. In preliminary studies, yeast containing no Pit-1 expression vectors showed little or no difference in /?-galactosidase expression from vectors containing zero, one, or four copies of a Pit-1binding site (data not shown). Similar findings were obtained when yeast carrying the pSEYC68GAL-Pit-1 vector were grown with glucose as the carbon source (Fig. 1). These findings suggest that Saccharomyces cerevisiae apparently contains no endogenous transcriptional activators that recognize the PRL Pit-1 1Pbinding site. All three of the Pit-1 expression vectors substantially activated expression of indicator plasmids containing four Pit-1-binding sites (Fig. 1, b and c). Little or no effect of Pit-1 expression vectors was observed for indicator plasmids containing a single Pit-1 -binding site (Fig. 1, b and c). As expected, activation mediated by production of Pit-1 from the GAL1 promoter was dependent on the use of galactose as the carbon source (Fig. 1b). Insertion of the Pit-1 coding sequence into any of the three expression vectors in the opposite orientation did not activate indicator gene expression (data not shown).
+ Glucose +Galactose pSEYC68GAL-Pit-1
pADH2-Pit-1
YCp88-Pit-1
Pit-1 Expression Vector
Fig. 1. Structure of Pit-1 Effector and Indicator Plasmids and Pit-1 Activation of Transcription in Saccharomyces cerevisiae The relative locations of functional elements in the Pit-1 effector plasmid and indicator plasmid are shown (a). An inducible Pit-1 expression vector was prepared which contained the GAL1 promoter upstream of the Pit-1 coding sequence, as indicated. The indicator plasmid contained either one or four copies of a synthetic Pit-1 1 P-binding site upstream from a CYC1-/acZ fusion gene. The relative locations of the origins of replication and selectable markers are also indicated. The abilities of the Pit-1 expression vectors containing either an inducible promoter (b) or constitutive promoters (c) to activate expression of the indicator plasmid were determined in yeast transformed with the indicated plasmids. Values are averages of duplicate determinations.
Similar Pit-1 domains are required for transcriptional activation in Saccharomyces cerevisiae and mammalian cells We compared the ability of expression vectors containing truncated Pit-1 coding sequences to activate indicator plasmid expression in both yeast and Rat-1 cells (Fig. 2). The results demonstrate that deletion of Pit-1 sequences has very similar effects in mammalian and yeast cells. Deletion of 44 residues from the aminoterminus of Pit-1 reduced frans-activation in yeast and Rat-1 cells. Deletion of 69 residues produced greater effects and substantially reduced the ability of Pit-1 to activate indicator plasmid expression, but did not abolish frans-activation. Again in both systems, removal of either 146 amino acids at the amino-terminus or removal of most of the POU-specific domain abolished frans-activation. The production of Pit-1 in transfected Rat-1 cells was also examined by gel mobility shift assays (Fig. 3). The mobility shift assays provided evi-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 05:45 For personal use only. No other uses without permission. . All rights reserved.
1241
Pit-1 Activation of Transcription in Yeast
Indicator Plasmid Expression (Relative Activity) • Yeast Cells, (J-galactosidase 10 20 30 40 50
Pit-1 Effector Plasmid Control 196
POU
213
272 291
HOMEO
4 H
8 12 16 Rat-1 Cells, luciferase
Fig. 2. Comparison of the Abilities of Expression Vectors Containing Wild-Type and Truncated Pit-1 Coding Sequences to Activate Transcription in Yeast and Mammalian Cells Truncated cDNAs encoding the indicated portions of Pit-1 were prepared and inserted in either YCp88 for transformation of yeast or in Rous sarcoma virus (RSV)-/3-globin for transfection of Rat-1 cells. The expression studies in yeast used the pCS104x1 P indicator plasmid. The Rat-1 cells were transfected by electroporation with 5 ^g of a 2.5PRL-Luc indicator plasmid (24) and 10 ng of an RSVPiti effector plasmid. The /3-galactososidase activity was determined in duplicate. Luciferase activity of triplicate samples was determined 24 h after transfection. To facilitate comparison of the data from yeast and Rat-1 cells, values are normalized to a value of 1.0 for the controls.
a>
Control
Wild Type
A N44
A N69
AN146
A147-201
£
f
Bound
Free
—
Fig. 3. Mobility shift analysis of extracts from Rat-1 cells transfected with Pit-1 expression vectors. Rat-1 cells were transfected with 10 ng of an RSV-Pit-1 expression vector encoding the indicated form of Pit-1. Cell extracts were prepared and assayed for DNA-binding activity by mobility shift assay using a radiolabeled 4x1 P Pit-1-binding site. Increasing concentrations of cell extract (3, 6, or 12 jug) were assayed as shown. The migration of the free DNA and slower migrating complexes is indicated at the left.
dence that the wild-type, AN44, and AN69 Pit-1 expression vectors produced similar levels of DNA-binding activity in the transfected cells. Indeed, although AN44 Pit-1 produced less frans-activation than the wild-type construct, this construct appeared to produce greater DNA-binding activity than the wild-type construct. Thus, the decreased frans-activating activity of the AN44 and AN69 constructs is apparently not due to decreased stability of the truncated proteins. Little or no DNAbinding activity was observed with the AN 146 and A147-201 constructs, consistent with previous reports that the POU-specific region of the Pit-1 is required for high affinity DNA binding (10). Mobility shift experiments
using Pit-1 produced in cell-free translation reactions confirmed that the wild-type, AN44, and AN69 Pit-1 proteins contain high affinity DNA-binding domains, but that the AN146 and A147-201 proteins do not (data not shown). Thus, the failure of AN146 and A147-201 Pit-1 to activate transcription is very likely due to destruction of DNA-binding activity, although we cannot rule out an effect on protein stability. Overall, the results of the deletion studies are consistent with previous studies in mammalian systems that mapped trans-acWvation domains of Pit-1 to an amino-terminal region and DNA-binding domains to the POU-specific and homeodomain regions (10,13). As observed in previous stud-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 05:45 For personal use only. No other uses without permission. . All rights reserved.
MOL ENDO-1991 1242
ies (10, 13), Pit-1 appears to contain multiple regions that are required for full transcriptional activation. The AN44 and AN69 constructs each produced only a portion of the activity obtained with the wild-type construct, and the levels differed substantially between these two constructs. The present findings demonstrate that these same regions of Pit-1 are required in yeast.
Vol 5 No. 9
a. Yeast Cells
p-Galactosidase Activity (units) • YCp88 • YCp88-Pit-1 .4
.8
1.2
1.6
CYC1 PRL -204 to + 34 I
Y/////////////A
- •
PRL +34 to -204 I
V///////////M
• « -
PRL-1714 to-1495 I
Y/////////////A
- •
PRL-1495 to-1714 I
Comparison of Pit-1-Binding Sites Required for Transcriptional Activation in Yeast and Mammalian Cells The initial experiments used an indicator plasmid containing either one or four copies of a Pit-1-binding site from the PRL gene. Two different regions of the PRL 5' flanking DNA have been shown to be important for transcriptional activation. One region is located immediately upstream from the TATA box, while a distal enhancer element is located about 1.5 kilobase pairs upstream from the transcription initiation site (22). Both the proximal and distal regions contain multiple Pit-1binding sites (4,23). Either of these regions can function as a transcriptional enhancer in mammalian cells (22, 24-26). Surprisingly, neither the 204 to 34 proximal region or the -1714 to -1495 distal enhancer of the PRL gene was able to confer Pit-1 responsiveness to the CYC1 promoter in yeast (Fig. 4). As in the preceding yeast studies, the 4x1 P-binding site was able to permit transcriptional activation in the presence of Pit-1. Analysis of similar constructs in GH3 cells, a mammalian cell line that expresses Pit1 (4, 5), demonstrated that indicator plasmids containing either the proximal PRL region or the 4x1 P-binding site could be activated (Fig. 4b).
DISCUSSION
These findings demonstrate that Pit-1 expression vectors can frans-activate expression of appropriate indicator plasmids in yeast. An important prerequisite for these studies was the finding that Saccharomyces cerevisiae does not contain an endogenous transcriptional activator that recognizes Pit-1-binding sites. Trans-activation in yeast was dependent on the presence of both a Pit-1 expression vector encoding a functional protein and an indicator plasmid containing Pit-1-binding sites. Thus, Pit-1, like a number of other mammalian transcription factors, can function in yeast. The finding that deletions of the Pit-1 coding sequence produced similar effects in yeast and mammalian cells demonstrates the requirement for similar functional domains of the protein in both systems. Furthermore, the fact that deletion of amino-terminal sequences of Pit-1 produced similar graded losses in transcriptional activation in both yeast and mammalian cells suggests the presence of activator subdomains that are required in both systems. This is particularly interesting in view of stud-
•
Yi Ding*, Wei Lu, Mark S. Roberson, W. Scott Moye-Rowley, and Richard A. Maurer Department of Physiology and Biophysics University of Iowa Iowa City, Iowa 52242
Pit-1 is a tissue-specific transcription factor which binds to specific DNA sequences within 5' flanking regions of the PRL and GH genes and activates the transcription of these genes. Previous studies have shown that expression of Pit-1 is necessary to activate transcription from the PRL or GH promoters in heterologous mammalian cells. In the present study the ability of Pit-1 expression vectors to activate expression of reporter genes in Saccharomyces cerevisiae was examined. The test system used Pit1 expression vectors and an indicator plasmid containing multiple copies of a Piti-binding site as a replacement for the upstream activator sequence of the CYC1 promoter. Significant activation of indicator plasmid expression was detected only in the presence of functional Pit-1 expression vectors. In both mammalian and yeast cells, amino-terminal deletions of the Pit-1 coding sequence produced similar and gradual loss of transcriptional activation. This finding indicates that similar or identical regions of Pit-1 are required for transcriptional activation in mammalian and yeast cells. Although synthetic DNA elements containing multiple copies of a single Pit1-binding site were sufficient to permit Pit-1-mediated transcriptional activation in both yeast and mammalian cells, DNA fragments representing the proximal region or distal enhancer region of the PRL gene were transcriptionally active only in mammalian cells. These studies establish the ability of Pit1 to stimulate transcription in the absence of other tissue-specific factors and provide a system for further genetic studies of Pit-1 structure/function relationships as well as evaluation of target sequences necessary for Pit-1 action. (Molecular Endocrinology 5: 1239-1245, 1991)
INTRODUCTION The pituitary-specific transcription factor Pit-1 is involved in the activation of specific genes within the 0888-8809/91/1239-1245$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society
anterior pituitary. Although there has been some controversy concerning the target genes that are activated by Pit-1 (1, 2), several studies suggest that Pit-1 can frans-activate both the PRL and GH genes (3-6). Pit-1 is a member of a family of transcription factors, designated POU factors, that contain two conserved domains (7). One of the conserved domains is similar to homeobox domains, found in a variety of developmentally important transcription factors (8, 9), while the other domain is specific to the POU factors. Both the POUspecific and the homeodomain are important for sequence-specific DNA binding of the POU factors (1012). The regions of Pit-1 that are required for transcriptional activation have also been explored. The aminoterminal portion of Pit-1 appears to contain several regions that are important for transcriptional activation of the GH and PRL genes (10,13). One experimental approach that has been important for demonstrating the role of Pit1 in transcriptional activation has involved transfection of heterologous mammalian cells with Pit-1 expression vectors (3-6). These studies have shown that Pit-1 is sufficient to activate the PRL and GH promoters when general transcription factors are provided by a heterologous cell. To examine the ability of Pit-1 to function in a nonmammalian cellular environment and to develop a system that permits genetic dissection of Pit-1 structure/function relationships, we have explored the ability of Pit-1 to activate transcription in yeast. The ability of several mammalian transcription factors to function in yeast (14-20) indicates the high degree of functional conservation of the transcriptional apparatus among eukaryotes. Even factors that are spatially restricted, such as the fushi tarazu homeobox-containing protein from Drosophila (21) have been demonstrated to function as transcriptional activators in yeast. While several transcription factors from higher eukaryotes have been shown to function in yeast, it is conceivable that at least some tissue-specific factors may require additional transcription factors that are either not present or are not functionally conserved in yeast. For instance, transcription factors that function as heterodimers containing a
1239
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 05:45 For personal use only. No other uses without permission. . All rights reserved.
Vol 5 No. 9
MOL ENDO-1991 1240
tissue-specific and a nontissue-specific subunit might function in heterologous mammalian cells, but not in yeast. Furthermore, POU domain proteins have not been found in nonmetazoan organisms. Thus, even though Pit-1 can activate transcription in heterologous mammalian cells, it is conceivable that Pit-1 might not function in yeast. In the present studies we demonstrate that Pit-1 can frans-activate gene expression in Saccharomyces cerevisiae. While several aspects of Pit-1dependent transcriptional activation were found to be similar between mammalian cells and Saccharomyces cerevisiae, a difference in DNA sequences that can support activation in the two systems was observed.
Pit-1 Binding Sites
a.
1X1P
EB 4x1 P CYC1 Promoter (minus UAS)
Pit-1 Coding Sequence
pSEYC68GAL-Pit-1 Effector Plasmid
pCS10-1x1Por4x1P Indicator Plasmid
RESULTS Pit-1 Expression in Saccharomyces cerevisiae and Activation of Gene Transcription To determine whether Pit-1 can function in yeast, three different Pit-1 expression plasmids under the control of heterologous promoters were constructed. A regulated expression vector was prepared by inserting the Pit-1 cDNA down-stream of the GAL1 promoter. Pit-1 expression using this vector should be dependent on the use of galactose as the carbon source. Constitutive expression vectors were also prepared using either the alcohol dehydrogenase promoter or the DED1 promoter carried on pADH2-Pit-1 and YCp88-Pit-1, respectively. The effects of Pit-1 expression were assessed by analysis of the expression of appropriate indicator plasmids. The indicator plasmids contained varying numbers of Pit-1-binding sites inserted up-stream of the yeast CYC1 promoter linked to lacZ. The CYC1 upstream activator sequence has been deleted from this construct, so that /3-galactosidase expression is dependent on binding of transcriptional activators to the inserted Pit-1-binding sites. In preliminary studies, yeast containing no Pit-1 expression vectors showed little or no difference in /?-galactosidase expression from vectors containing zero, one, or four copies of a Pit-1binding site (data not shown). Similar findings were obtained when yeast carrying the pSEYC68GAL-Pit-1 vector were grown with glucose as the carbon source (Fig. 1). These findings suggest that Saccharomyces cerevisiae apparently contains no endogenous transcriptional activators that recognize the PRL Pit-1 1Pbinding site. All three of the Pit-1 expression vectors substantially activated expression of indicator plasmids containing four Pit-1-binding sites (Fig. 1, b and c). Little or no effect of Pit-1 expression vectors was observed for indicator plasmids containing a single Pit-1 -binding site (Fig. 1, b and c). As expected, activation mediated by production of Pit-1 from the GAL1 promoter was dependent on the use of galactose as the carbon source (Fig. 1b). Insertion of the Pit-1 coding sequence into any of the three expression vectors in the opposite orientation did not activate indicator gene expression (data not shown).
+ Glucose +Galactose pSEYC68GAL-Pit-1
pADH2-Pit-1
YCp88-Pit-1
Pit-1 Expression Vector
Fig. 1. Structure of Pit-1 Effector and Indicator Plasmids and Pit-1 Activation of Transcription in Saccharomyces cerevisiae The relative locations of functional elements in the Pit-1 effector plasmid and indicator plasmid are shown (a). An inducible Pit-1 expression vector was prepared which contained the GAL1 promoter upstream of the Pit-1 coding sequence, as indicated. The indicator plasmid contained either one or four copies of a synthetic Pit-1 1 P-binding site upstream from a CYC1-/acZ fusion gene. The relative locations of the origins of replication and selectable markers are also indicated. The abilities of the Pit-1 expression vectors containing either an inducible promoter (b) or constitutive promoters (c) to activate expression of the indicator plasmid were determined in yeast transformed with the indicated plasmids. Values are averages of duplicate determinations.
Similar Pit-1 domains are required for transcriptional activation in Saccharomyces cerevisiae and mammalian cells We compared the ability of expression vectors containing truncated Pit-1 coding sequences to activate indicator plasmid expression in both yeast and Rat-1 cells (Fig. 2). The results demonstrate that deletion of Pit-1 sequences has very similar effects in mammalian and yeast cells. Deletion of 44 residues from the aminoterminus of Pit-1 reduced frans-activation in yeast and Rat-1 cells. Deletion of 69 residues produced greater effects and substantially reduced the ability of Pit-1 to activate indicator plasmid expression, but did not abolish frans-activation. Again in both systems, removal of either 146 amino acids at the amino-terminus or removal of most of the POU-specific domain abolished frans-activation. The production of Pit-1 in transfected Rat-1 cells was also examined by gel mobility shift assays (Fig. 3). The mobility shift assays provided evi-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 05:45 For personal use only. No other uses without permission. . All rights reserved.
1241
Pit-1 Activation of Transcription in Yeast
Indicator Plasmid Expression (Relative Activity) • Yeast Cells, (J-galactosidase 10 20 30 40 50
Pit-1 Effector Plasmid Control 196
POU
213
272 291
HOMEO
4 H
8 12 16 Rat-1 Cells, luciferase
Fig. 2. Comparison of the Abilities of Expression Vectors Containing Wild-Type and Truncated Pit-1 Coding Sequences to Activate Transcription in Yeast and Mammalian Cells Truncated cDNAs encoding the indicated portions of Pit-1 were prepared and inserted in either YCp88 for transformation of yeast or in Rous sarcoma virus (RSV)-/3-globin for transfection of Rat-1 cells. The expression studies in yeast used the pCS104x1 P indicator plasmid. The Rat-1 cells were transfected by electroporation with 5 ^g of a 2.5PRL-Luc indicator plasmid (24) and 10 ng of an RSVPiti effector plasmid. The /3-galactososidase activity was determined in duplicate. Luciferase activity of triplicate samples was determined 24 h after transfection. To facilitate comparison of the data from yeast and Rat-1 cells, values are normalized to a value of 1.0 for the controls.
a>
Control
Wild Type
A N44
A N69
AN146
A147-201
£
f
Bound
Free
—
Fig. 3. Mobility shift analysis of extracts from Rat-1 cells transfected with Pit-1 expression vectors. Rat-1 cells were transfected with 10 ng of an RSV-Pit-1 expression vector encoding the indicated form of Pit-1. Cell extracts were prepared and assayed for DNA-binding activity by mobility shift assay using a radiolabeled 4x1 P Pit-1-binding site. Increasing concentrations of cell extract (3, 6, or 12 jug) were assayed as shown. The migration of the free DNA and slower migrating complexes is indicated at the left.
dence that the wild-type, AN44, and AN69 Pit-1 expression vectors produced similar levels of DNA-binding activity in the transfected cells. Indeed, although AN44 Pit-1 produced less frans-activation than the wild-type construct, this construct appeared to produce greater DNA-binding activity than the wild-type construct. Thus, the decreased frans-activating activity of the AN44 and AN69 constructs is apparently not due to decreased stability of the truncated proteins. Little or no DNAbinding activity was observed with the AN 146 and A147-201 constructs, consistent with previous reports that the POU-specific region of the Pit-1 is required for high affinity DNA binding (10). Mobility shift experiments
using Pit-1 produced in cell-free translation reactions confirmed that the wild-type, AN44, and AN69 Pit-1 proteins contain high affinity DNA-binding domains, but that the AN146 and A147-201 proteins do not (data not shown). Thus, the failure of AN146 and A147-201 Pit-1 to activate transcription is very likely due to destruction of DNA-binding activity, although we cannot rule out an effect on protein stability. Overall, the results of the deletion studies are consistent with previous studies in mammalian systems that mapped trans-acWvation domains of Pit-1 to an amino-terminal region and DNA-binding domains to the POU-specific and homeodomain regions (10,13). As observed in previous stud-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 05:45 For personal use only. No other uses without permission. . All rights reserved.
MOL ENDO-1991 1242
ies (10, 13), Pit-1 appears to contain multiple regions that are required for full transcriptional activation. The AN44 and AN69 constructs each produced only a portion of the activity obtained with the wild-type construct, and the levels differed substantially between these two constructs. The present findings demonstrate that these same regions of Pit-1 are required in yeast.
Vol 5 No. 9
a. Yeast Cells
p-Galactosidase Activity (units) • YCp88 • YCp88-Pit-1 .4
.8
1.2
1.6
CYC1 PRL -204 to + 34 I
Y/////////////A
- •
PRL +34 to -204 I
V///////////M
• « -
PRL-1714 to-1495 I
Y/////////////A
- •
PRL-1495 to-1714 I
Comparison of Pit-1-Binding Sites Required for Transcriptional Activation in Yeast and Mammalian Cells The initial experiments used an indicator plasmid containing either one or four copies of a Pit-1-binding site from the PRL gene. Two different regions of the PRL 5' flanking DNA have been shown to be important for transcriptional activation. One region is located immediately upstream from the TATA box, while a distal enhancer element is located about 1.5 kilobase pairs upstream from the transcription initiation site (22). Both the proximal and distal regions contain multiple Pit-1binding sites (4,23). Either of these regions can function as a transcriptional enhancer in mammalian cells (22, 24-26). Surprisingly, neither the 204 to 34 proximal region or the -1714 to -1495 distal enhancer of the PRL gene was able to confer Pit-1 responsiveness to the CYC1 promoter in yeast (Fig. 4). As in the preceding yeast studies, the 4x1 P-binding site was able to permit transcriptional activation in the presence of Pit-1. Analysis of similar constructs in GH3 cells, a mammalian cell line that expresses Pit1 (4, 5), demonstrated that indicator plasmids containing either the proximal PRL region or the 4x1 P-binding site could be activated (Fig. 4b).
DISCUSSION
These findings demonstrate that Pit-1 expression vectors can frans-activate expression of appropriate indicator plasmids in yeast. An important prerequisite for these studies was the finding that Saccharomyces cerevisiae does not contain an endogenous transcriptional activator that recognizes Pit-1-binding sites. Trans-activation in yeast was dependent on the presence of both a Pit-1 expression vector encoding a functional protein and an indicator plasmid containing Pit-1-binding sites. Thus, Pit-1, like a number of other mammalian transcription factors, can function in yeast. The finding that deletions of the Pit-1 coding sequence produced similar effects in yeast and mammalian cells demonstrates the requirement for similar functional domains of the protein in both systems. Furthermore, the fact that deletion of amino-terminal sequences of Pit-1 produced similar graded losses in transcriptional activation in both yeast and mammalian cells suggests the presence of activator subdomains that are required in both systems. This is particularly interesting in view of stud-
•
