| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Pathology, Northwestern University, Chicago, Illinois 60611
Address all correspondence and requests for reprints to: Charles Clevenger, Department of Pathology, Northwestern University, Lurie 4-107, 303 East Superior Street, Chicago, Illinois 60611. E-mail: clevenger{at}northwestern.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
S1 PRLrs dimerized in a ligand-independent manner. To determine the domain(s) involved in PRLr ligand-independent dimerization, we generated PRLr constructs as follows: (1) the TM-ICD, which consisted of the TM domain and the intracellular domain (ICD) but lacked the extracellular domain (ECD), and (2) the ECD-TM, which consisted of the TM domain and the ECD but lacked the ICD. These constructs dimerized in a ligand-independent manner in mammalian cells, implicating a significant role for the TM domain in this process. These truncated PRLrs were functionally inert alone or in combination in cells lacking the PRLr. However, when introduced into cells containing endogenous PRLr, the ECD-TM inhibited human PRLr signaling, whereas the TM-ICD potentiated human PRLr signaling. These studies indicate that the ECD-TM and the TM-ICD are capable of modulating PRLr function. We also demonstrated an endogenous TM-ICD in T47D cells, suggesting that these findings are relevant to PRL-signaling pathways in breast cancer. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
There are presently six known isoforms of the hPRLr. The long hPRLr isoform is a classic type I transmembrane (TM) receptor and a member of the larger family of cytokine receptors (3, 20). The extracellular (ligand engagement) domain (ECD) consists of two sub domains (S1 and S2), each structurally configured as a ß-sandwich. The intracellular (signaling) domain (ICD) of the hPRLr contains a region of proximal homology to other cytokine receptors, i.e. the Box1/Variable Box/Box 2/X box, as well as a unique C-terminal tail. The intermediate hPRLr isoform is generated from an alternative splice within the last hPRLr exon, resulting in a deletion of all coding sequences C-terminal to the X-Box and the addition of a novel 13 amino acid sequence, thereby giving the intermediate hPRLr isoform different signaling properties than the long hPRLr isoform (21). The S1 hPRLr isoform also represents a splicing variant, which results in a hPRLr that lacks the first ß-sandwich ("S1") domain of the hPRLr ECD (7); thus, the affinity of S1 hPRLr homodimers for PRL is 7-fold lower than that observed for long and intermediate hPRLr isoforms (22). Two short forms of the hPRLr have been identified; the S1a isoform contains both Box1 and 2 elements, in addition to a C-terminal unique 39 amino acid sequence, whereas the S1b isoform contains only Box1(23). In most studies (23, 24), but not all (25), the short forms appear to down-regulate PRL-induced signaling.
The PRL binding protein (PRLBP) represents the freely circulating ECD of the hPRLr. Given the absence of corresponding mRNA, the PRLBP is thought to arise from a proteolytic event rather than a splicing event (26, 27). Several studies have reported the freely circulating GH binding protein (GHBP) (28, 29, 30) via a proteolytic event. These events occur via a characterized pathway and result in the generation of the GHBP as well as a remnant GHr containing the TM domain and the ICD (31, 32). The events that generate the PRLBP are unknown; however, data presented herein are the first to suggest that a similar remnant of the PRLr, namely the TM-ICD, is generated via cleavage of the PRLr ECD.
PRL binds to the PRLr via 2 interaction sites, referred to as binding sites 1 and 2. The accepted view has been that the initial event in PRL-induced signaling is the binding of PRL to one hPRLr molecule via binding site 1, and this 1:1 complex recruits a second PRLr molecule via binding site 2 of PRL, inducing sequential dimerization of the hPRLr (33, 34, 35). However, other laboratories have reported that cytokine receptors such as those for GH and erythropoietin (EPO) exist as dimers in the absence of ligand (18, 36, 37, 38, 39, 40). These studies indicate that ligand-independent predimerization of both receptors occurs via the TM domain (37, 41, 42). This predimerization of the receptor, however, is insufficient to trigger signaling in the absence of ligand.
Studies presented here confirm the ligand-independent dimerization of endogenous long and intermediate PRLr isoforms as well as transfected differentially epitope-tagged long, intermediate and S1 hPRLr isoforms. Data are presented that demonstrate that the TM domain is sufficient for this process; however, our findings suggest that this interaction is strengthened by both the ECD and the ICD. Although functional studies demonstrate that ECD-TM and TM-ICD constructs are functionally inert in luciferase assays, transfection of the TM-ICD construct potentiated endogenous hPRLr signaling in T47D breast cancer cells, suggesting that a hPRLr lacking the ICD is capable of facilitating the transduction of PRL signal. Because the TM-ICD PRLr fragment appeared to occur endogenously in T47D cells via shedding of the ECD, these findings could be physiologically relevant to PRL-signaling pathways in breast cancer.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, these studies used as bait a 102-amino acid fragment near the C terminus of the long isoform, or a 24-amino acid fragment of the intermediate isoform. The fragment of the long hPRLr included a central region C-terminal to the Box1/variable box/Box 2 domains of the hPRLr that are important for Vav1 and Tec binding; interestingly, this region contains a conserved bipartite W(L/P)LPQ motif (43). The fragment of the intermediate hPRLr included the unique 13 amino acids of the C-terminal tail of the intermediate receptor. Thus, we chose these fragments for yeast two-hybrid analysis to determine proteins that interact with these two unique regions in the long and intermediate hPRLr isoforms. It was necessary to use relatively small intracellular fragments of the long and intermediate isoforms given that full-length receptors/large fragments led to transactivation in the yeast two-hybrid assay. When using both of these small fragments as bait to screen a cDNA library from T47D breast cancer cells, 5% of the clones obtained were the full-length long hPRLr isoform (Fig. 1B). This indicated that long-long and long-intermediate dimers formed in yeast cells in the absence of PRL.
|
(top panel), the long PRLr isoform immunoprecipitated with the intermediate PRLr isoform in the presence or absence of PRL, indicating ligand-independent complex formation in T47D cells. Interestingly, the extent of association between these isoforms did not change in the presence of FBS (lane 1) or in the presence of PRL (lane 2). This is a specific interaction, as shown by the IgG control (lane 4). Parallel studies measured the amount of long hPRLr immunoprecipitated under the same conditions, and these studies indicated that roughly 30–40% of long hPRLr immunoprecipitates with intermediate hPRLr (Fig. 2A, bottom panel, lane 1) in the absence of exogenous ligand. This calculation is based on the observation that PRLr is completely immunoprecipitated in this assay. This was demonstrated by subjecting the supernatant from the first immunoprecipitation procedure to a second immunoprecipitation with the PRLr-ECD antibody; as shown in Fig. 2A (bottom panel, lane 2), no PRLr was detected in this supernatant fraction.
|
, and as previously documented (44), T47D cells express small amounts of endogenous PRL when maintained for 48 h in defined medium, whereas HepG2 cells do not produce detectable amounts of PRL under the same conditions. Using recombinant hPRL as a standard, we determined that T47D cell lysates contain approximately 5 ng PRL/1.2 million cells (data not shown). Figure 2C shows that T47D cells and HepG2 cells express comparable levels of long and intermediate PRLr isoforms. To determine whether PRLr dimerization occurs in HepG2 cells in the absence of PRL, immunoprecipitations were performed in HepG2 cells as described for T47D cells, and the results are shown in Fig. 2D. These studies confirmed the ligand-independent association between long and intermediate isoforms in HepG2 cells in the absence of PRL. Similar to T47D cells, approximately 30–40% of the long isoform coimmunoprecipitated with the intermediate isoform in the absence of exogenous ligand.
Long, Intermediate and S1 hPRLrs Form Ligand-Independent Homo/Heterodimers in Transfected Systems
To detect ligand-independent homodimerization of long or intermediate hPRLr, we extended these studies to transfected mammalian cells to enhance detection of specific predimerized complexes. COS7 monkey kidney cells lack endogenous PRL and readily express PRLr expression constructs, and therefore serve as a useful model to study the ligand-independent interaction between specific transfected PRLr isoforms with enhanced sensitivity. In addition, COS7 cells transfected with PRLr show PRL signaling such as PRL-induced phosphorylation of Stat5a (data not shown), demonstrating that PRL stimulation is effective via the transfected receptor in these cells. Given that some of the PRLr isoforms lack specific domains as a consequence of splicing (i.e. intermediate, S1), the data obtained from these studies allowed us to examine whether the domains that are deleted in these receptor constructs are necessary for ligand-independent association. To measure ligand-independent dimerization of differentially epitope-tagged hPRLr isoforms, COS7 cells were cotransfected with hPRLr isoforms with C-terminal Myc or V5 tags. Cell lysates were then immunoprecipitated with an anti-Myc antibody and then blotted with an anti-V5 antibody. These studies showed that the hPRLr formed long-long homodimers in a ligand-independent manner in COS7 cells (Fig. 3A, top panel). As before, the membrane was stripped and reprobed with an anti-Myc antibody to confirm both expression (lysate lane) and immunoprecipitation (IP lanes) of Myc-tagged long PRLr (Fig. 3A, middle panel). These studies were extended to the intermediate hPRLr, indicating that the transfected intermediate hPRLr also homodimerized in a ligand-independent manner in COS7 cells (Fig. 3B, top panel); stripping and reprobing of this membrane confirmed both expression and immunoprecipitation of Myc-tagged intermediate hPRLr (Fig. 3B, middle panel). To detect ligand-independent heterodimerization of long and intermediate hPRLr, COS7 cells were cotransfected with differentially epitope tagged long and intermediate hPRLrs. These studies showed that the hPRLr formed long-intermediate heterodimers in a ligand-independent manner in COS7 cells (Fig. 3C, top panel). This membrane was stripped and reprobed with an anti-Myc antibody to confirm both expression and immunoprecipitation of Myc-tagged long PRLr (Fig. 3C, bottom panel). Quantitation of bands indicated that, on average, approximately 30–35% of transfected long and intermediate PRLr isoforms were hetero/homodimerized in a ligand-independent manner. To confirm that immunoprecipitation of epitope-tagged receptors was not due to stickiness of epitope tags, we performed controls in which COS7 cells were cotransfected with Myc-tagged long hPRLr and V5-tagged Cyclophilin B (CypB), and performed immunoprecipitations as described above. As expected, CypB and long hPRLr did not coimmunoprecipitate, indicating that detection of ligand-independent dimerization is not due to sticky epitope tagged forms of the hPRLr (Fig. 3D).
|
, A and B, top panel). These shorter forms were not detectable with long-Myc tagged PRLr (Fig. 3A) and were only slightly detectable with intermediate-Myc tagged PRLr (Fig. 3B), an effect that is probably due to antibody sensitivity. Due to the fact that the V5 tag is C-terminal, this suggested that a product that is cleaved in the ECD is immunoprecipitating with the Myc-tagged isoforms. To test this hypothesis, parallel samples were prepared as described above, and immunoblotted with an anti-PRLr-ECD antibody. As shown in Fig. 3, A and B (bottom panels), the anti-PRLr-ECD antibody detected only full-length long and intermediate hPRLr, suggesting that these smaller V5-tagged bands correspond to epitope tagged PRLr lacking a significant portion, if not all, of its ECD. Because these smaller V5-tagged PRLrs efficiently immunoprecipitated with the Myc-tagged PRLr, this suggests that the ECD is not essential for ligand-independent hPRLr dimerization. To determine whether cleavage of the ECD also occurs with endogenous hPRLr, parallel lysates from T47D breast cancer cells were immunoblotted with the anti-PRLr-ECD antibody or with an anti-PRLr-ICD antibody. The anti-PRLr-ICD epitope corresponds to amino acids 323–622 of the hPRLr and therefore reacts only with long and S1 hPRLrs. As shown in Fig. 3E, the anti-PRLr-ECD antibody primarily immunoreacted with the long and intermediate hPRLr isoforms, as well as the PRLBP (left panel). The anti-PRLr-ICD antibody immunoreacted with the long and S1 hPRLr isoforms as expected; however, a strong band was also detected at approximately 55–60 kDa, suggesting the presence of long hPRLr or S1 hPRLr that has been cleaved in the ECD (ECD-PRLr, Fig. 3E, right panel). These data indicate that the cleavage of the ECD occurs with endogenous hPRLr and is not an artifact of transfection.
To further address the role of the ECD in ligand-independent dimerization, we used the S1 hPRLr isoform, which lacks the extracellular S1 domain. COS7 cells were cotransfected with differentially epitope tagged long and S1 hPRLrs (Fig. 4A), differentially tagged intermediate and S1 hPRLrs (Fig. 4B), or differentially tagged S1 hPRLrs (Fig. 4C). These studies showed that both the long and intermediate hPRLrs heterodimerized with S1 hPRLr in a ligand-independent manner in COS7 cells (Fig. 4, A and B). These studies also showed that the S1 hPRLr homodimerized in a ligand-independent manner in COS7 cells (Fig. 4C). These data are consistent with Fig. 3, A–C, which suggest that transfected long and intermediate hPRLrs with cleaved ECDs dimerize in a ligand-independent manner. The ability of the S1 hPRLr to form ligand-independent heterodimers with long and intermediate hPRLrs, in addition to the ability of the S1 hPRLr to homodimerize, further suggested that the S1 domain is not critical for this process.
|
. To determine whether the hPRLr can dimerize in a ligand-independent manner if one of the PRLrs lacks the ECD, cells were cotransfected with differentially epitope tagged long hPRLr and TM-ICD hPRLr (Fig. 5B, lanes 1–4). These studies showed that the long and TM-ICD PRLrs heterodimerized in a ligand-independent manner in COS7 cells; however, this interaction was weaker than the interactions seen with the other hPRLr isoforms in both the presence and absence of PRL.
|
. To determine whether the hPRLr can dimerize in a ligand-independent manner if one of the PRLrs lacks the ICD, COS7 cells were cotransfected with differentially epitope-tagged ECD-TM hPRLr and long hPRLr (Fig. 5B, lanes 9–12). As shown, ECD-TM PRLr and long hPRLr formed ligand-independent heterodimers in the presence and absence of PRL in COS7 cells, indicating that the ICD is not necessary for ligand-independent dimerization. Membranes were stripped and reprobed with an anti-Myc antibody to confirm expression and immunoprecipitation of Myc-tagged TM-ICD mutant PRLr (Fig. 5B, lower panels).
Previous studies have indicated that the TM domain is necessary for ligand-independent dimerization of other cytokine receptors (37, 41, 42). To determine whether the TM domain is sufficient for ligand-independent dimerization of the hPRLr, COS7 cells were cotransfected with differentially epitope tagged ECD-TM hPRLr and TM-ICD hPRLr (Fig. 5B, lanes 5–8). As shown, these constructs formed ligand-independent dimers in COS7 cells, indicating that the TM domain is sufficient for inducing this interaction. However, this interaction is weaker than the interactions between hPRLr isoforms in both the presence and absence of PRL. Whereas approximately 30–50% of the other transfected PRLr isoforms are dimerized in a ligand-independent manner, only 12–15% of the mutant PRLrs are dimerized in the absence of ligand. This calculation is based on direct comparisons of coimmunoprecipitation experiments detecting dimerization between the long-long isoforms and the long-TM-ICD and long-ECD-TM isoforms (data not shown). These data suggest that although the TM domain is sufficient to induce ligand-independent dimerization of the PRLr, other domains may be involved in stabilizing this interaction.
ECD-TM and TM-ICD Mutant hPRLrs Demonstrate Differential Signaling Properties in Functional Assays
Luciferase assays were used to determine the functional consequences of ECD-TM and TM-ICD mutant PRLrs on wild-type hPRLr signaling. For initial assays, we used Chinese hamster ovary (CHO) cells because they do not express PRLr and therefore do not have endogenous PRL signaling pathways in the absence of transfected hPRLr. To examine the ability of the TM-ICD and ECD-TM mutant PRLrs to activate Stat5a, CHO cells were cotransfected with equal amounts of ECD-TM or TM-ICD mutant PRLrs. As expected, ECD-TM or TM-ICD transfected alone or together did not induce Stat5a transactivation in CHO cells (data not shown). We extended these studies to T47D breast cancer cells to determine the effect of transfected ECD-TM and TM-ICD mutants on endogenous hPRLr signaling. Figure 6A shows that ECD-TM inhibited endogenous hPRLr signaling, consistent with previous reports demonstrating that C-terminal truncated PRLrs can act as inhibitors of wild-type PRLr signaling (23, 24, 45, 46). Surprisingly, a dose response study demonstrated that increasing amounts of TM-ICD resulted in a statistically significant increase in Stat5a transactivation (Fig. 6B), suggesting that the TM-ICD, and possibly the heterodimers formed with the endogenous hPRLr, are of functional importance in T47D cells.
|
). Studies were also performed with cells rested in serum-depleted medium and dosed with PRL to ensure levels did not change under same conditions used in the luciferase assay (data not shown). These studies indicated that the TM-ICD PRLr did not appear to affect wild-type receptor steady-state levels. To ensure that the TM-ICD PRLr was expressed in the membrane fraction of T47D cells, T47D cells transfected with TM-ICD PRLr were fractionated into membrane and cytoplasmic compartments and lysates were immunoblotted with an anti-Myc antibody. These studies demonstrated that the majority of the TM-ICD was expressed in the membrane fraction (data not shown). To assess whether the TM-ICD was phosphorylated in response to PRL stimulation, parallel anti-PRLr-ICD immunoprecipitates were probed by antiphosphotyrosine immunoblot analysis (Fig. 7B), demonstrating a 70% increase in TM-ICD phosphorylation and a 20% increase in endogenous hPRLr phosphorylation (Fig. 7B, lower right and lower left panels, respectively). These studies suggest that the TM-ICD is targeted by the tyrosine kinases involved in PRL signaling.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Crystal structure studies of the GH binding protein (GHBP) showed that GHBP existed as a dimer with GH (50), but as a monomer with GH antagonists (13, 51), indicating that GH antagonists exerted their effects by preventing the functional dimerization of GH receptor. However, other studies using full-length GHr (i.e. containing the TM domain) as opposed to GHBP have shown that GH antagonists appear to bind to a preformed dimer (18, 36). A further study confirmed predimerization of GHr and suggested that this process occurs in the endoplastic reticulum via the transmembrane domain (37). A recent paper has suggested that GH activates a preformed dimer by transmission of rotational torque through the TM resulting in rotation of the GHr ICDs (40). However, because these studies solely involved overexpression of transfected GHr constructs, these authors raised the concern that the observed phenomenon may not occur at physiological concentrations of GHr. Studies presented here demonstrate ligand-independent dimerization of endogenous long and intermediate hPRLr isoforms in T47D and HepG2 cells, indicating that such dimerization of the hPRLr occurs at physiologic concentrations and is not an artifact of transfection. These data also demonstrate that the addition of PRL to PRL-starved cells does not result in a significant increase in dimer formation, suggesting that a sizable portion of the hPRLr is predimerized in these cell lines. Similar to studies of endogenous hPRLr, we observed no increase in dimer formation upon addition of PRL in COS7 cells transfected with long, intermediate, and S1 hPRLr isoforms. In addition, a recent study demonstrated ligand-independent homo- and heterodimerization of the transfected long and two short isoforms by bioluminescence resonance energy transfer (52). These results are in contrast to previous studies demonstrating a lack of fluorescence resonance energy transfer signal between PRLr and GHr heterodimers in the absence of ligand (53), and studies demonstrating a lack of bioluminescence resonance energy transfer signal between homo- and heterodimers of long and the two short PRLr isoforms in the absence of ligand (25). Both of these latter studies demonstrated increased signal in response to ligand, indicating that ligand induced a decrease in the distance between the ICDs of the two receptors. However, these techniques do not differentiate between receptors whose ICDs undergo a conformational change upon ligand binding, or receptors that are predimerized in the absence of ligand.
The presence of smaller products in COS7 transfectant cell lysates that immunoreact with the C-terminal PRLr antibody but not with the N-terminal PRLr antibody indicated generation of products that are lacking all or part of the ECD. These smaller products formed pre-dimers with full-length products, suggesting that the entire ECD is not necessary for initiation of ligand-independent dimerization. The generation of this fragment lacking the ECD was not an artifact of transfection because a fragment of similar size was detected in T47D cell lysates when using an antibody specific for the C-terminal end of the hPRLr (Fig. 3E). A precedent exists for such cleavage in the GHr, which is cleaved to yield the GHBP and a fragment consisting of the TM domain and the ICD (31, 32). The GHBP results from the cleavage of GH receptor (GHr) by TNF-
converting enzyme (TACE), a metzincin metalloprotease that catalyzes the proteolytic cleavage of cell surface molecules in response to phorbol 12-myristate 13-acetate (31). Further studies suggested that the metalloprotease GHr remnant may have signaling properties (32). A similar mechanism may serve to generate the freely circulating ECD of the PRLr, the PRLBP (26), although it is probable that other enzymatic processes may also contribute (27).
To further assess the importance of the ECD and the ICD in ligand-independent dimerization, we generated PRLr constructs that consisted of the TM domain and the ICD but lacked the ECD (TM-ICD), and a construct that consisted of the TM domain and the ECD but lacked the ICD (ECD-TM). The ECD-TM formed ligand-independent dimers with the long hPRLr in COS7 cells, suggesting that it is not necessary to have two ICDs for the initiation of ligand-independent dimerization of the hPRLr. Similarly, the TM-ICD formed ligand-independent dimers with the long hPRLr in COS7 cells, indicating that two ECDs are also not necessary for the initiation of ligand-independent dimerization of the hPRLr. Further studies showed that the TM-ICD formed a ligand-independent dimer with ECD-TM (Fig. 5), confirming that the TM domain facilitates ligand-independent dimerization of the hPRLr. However, the association between the TM-ICD and ECD-TM together or with the long hPRLr isoform were less robust, suggesting that although the TM domain is necessary for this association, the interactions between the two ECDs and two ICDs strengthen this interaction. Consistent with this idea, there was a small but reproducible increase in ligand-independent dimer formation between long hPRLr and the TM-ICD and ECD-TM mutants compared with ECD-TM and TM-ICD mutants together. This is also consistent with yeast two-hybrid data (Fig. 1) that indicated that intracellular fragments of both the long and intermediate hPRLr isoforms formed ligand-independent dimers with full-length long PRLr in yeast cells. These data demonstrating a role for the TM domain in the ligand-independent dimerization of the hPRLr might explain why a previous study demonstrated that recombinant soluble PRLr-ECD (PRLBP) does not predimerize but rather forms an unstable, transient homodimer upon PRL binding (34). However, the data presented here argues that a significant fraction of the full-length PRLr is stably predimerized via the TM domain.
Functional studies showed that although the TM domain is sufficient to induce ligand-independent dimerization of PRLr constructs that lack the ECD and the ICD, this interaction was not sufficient to transduce PRL signaling in dimeric pairs containing only one intact ICD. This was demonstrated by the fact that CHO cells cotransfected with TM-ICD and ECD-TM mutant PRLr constructs either separately or in combination, failed to transactivate Stat5a in luciferase assays. Additional studies used T47D cells to determine the effect of these mutant PRLrs on endogenous hPRLr signaling. As expected, the ECD-TM PRLr mutant that lacks the ICD inhibited PRL-signaling in T47D cells. This is consistent with previous studies demonstrating that the short forms of the hPRLr (23, 24) and the rat PRLr (45), as well as truncated hPRLr (46, 54) inhibit PRL-signaling via the long PRLr isoform. Taken together, these data indicate the need for two functional ICDs in PRLr homo/heterodimers to transduce PRL signal. Unexpectedly, increasing amounts of the TM-ICD mutant PRLr in T47D cells resulted in a small but significant increase in PRL signaling via endogenous hPRLr in T47D cells (Fig. 6), demonstrating that although this mutant lacked the ECD, it was still able to potentiate Stat5a transactivation. The observation that increased levels of transfected TM-ICD was required to see a small, but significant, increase in Stat5a transactivation would be consistent with the fact that dimers formed between long hPRLr and TM-ICD mutant PRLr were weaker than dimers formed between other hPRLr isoforms. To confirm that this increase in Stat5a transactivation was due specifically to PRL signaling via the TM-ICD and not due to a nonspecific effect of transfection, the proper localization and PRL-induced phosphorylation was confirmed in T47D cells (data not shown and Fig. 7B). One interpretation of our findings is that the TM-ICD contributes to PRL signal transduction. This may indicate that one ECD is sufficient to bind PRL and induce the necessary ICD conformational change in a wild-type:TM-ICD heterodimer that allows transduction of the PRL signal. Alternative explanations of these findings include: 1) the TM-ICD fragment may prevent degradation of endogenous hPRLr in T47D cells, therefore increasing PRL signaling; 2) the TM-ICD fragment may act as a decoy for inhibitors/phosphatases that act upon wild-type PRLr; 3) the TM-ICD fragment may act directly on Stat5a transcriptional activity in the nucleus, as we have previously observed that wild-type hPRLr is translocated to the nucleus where it potentiated Stat5a transactivation (our unpublished observations); or 4) the TM-ICD fragment may function in a multimeric complex of PRLrs. However, measurement of the amount of wild-type hPRLr in T47D cells transfected with either TM-ICD or vector found no consistent increase in wild-type hPRLr in these cells (Fig. 7A), suggesting that the mechanism of increased Stat5a transactivation is not due to increased steady-state levels of wild-type hPRLr. In addition, our observation of the PRL-inducible phosphorylation of the TM-ICD in T47D cells suggested that the TM-ICD is closely involved in the PRL signaling pathway, and therefore we feel that our initial interpretation is most likely.
This finding could explain previous studies demonstrating the ineffectiveness of the G129R PRL antagonist. The G129R PRL was developed as a potential antagonist because it contained a normal binding site 1, but a mutation in binding site 2, which would thereby prevent the functional dimerization of the PRLr (25, 55). However, G129R acted as a partial antagonist in Nb2 with weak agonist properties (56). One explanation of these data has been that the site 2 of G129R was not sufficiently altered to prevent PRLr dimerization and therefore did not inhibit signaling. However, another explanation arising from the findings presented here is that the binding of G129R to only one of the PRLrs in the PRLr pre-dimer is sufficient to induce functional conformational changes in the ICD in either wild-type or TM-ICD PRLr, resulting in PRL signaling.
Taken together, these data suggest a functional role for the TM-ICD PRLr. Because generation of the TM-ICD hPRLr via cleavage of the ECD appears to occur endogenously in T47D cells (Fig. 3D; Ref. 27), this finding could be physiologically relevant to PRL-signaling pathways in breast cancer, and further guide the development of PRL antagonists.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Generation of cDNA Constructs
The hPRLr V5 and Myc C-terminal tagged S1 isoform, and intermediate isoform, and full-length long isoform of the hPRLr have been previously described (7, 21, 43), respectively; and the V5-tagged cypB construct has been previously described (57). Using the full-length hPRLr as template, TM-ICD mutant hPRLr was generated using the following primers: forward, 5'-ATAAGAATGCGGCCGCTAAACTATCACCATGATGGGATGGAGCTGTATCATCCTCTTCTTGGTAGCAACAGCTACAACAACGTGGATCTCGTG-3', and reverse, 5'-CTAGTCTAGACTAGGGTGAAAGGAGTGTGT-3'. The forward primer contains the NotI restriction site, Kozak initiation sequence, an ER leader sequence, and the hPRLr sequence beginning at the TM domain; and the reverse primer contains the XbaI restriction site and the 3' end of the full-length long hPRLr. These primers rendered a fragment containing an ER leader sequence, the TM domain and the full ICD of the long hPRLr. This fragment was isolated and inserted into pCDNA3.1-Myc/His (Invitrogen, Carlsbad, CA) using the NotI/XbaI restriction sites. Using the full-length hPRLr as template, ECD-TM mutant hPRLr was generated using the following primers: forward, 5'-TGGTGGAATTCTCCACATGACAACCGTGTGG-3', reverse, 5'-GCGGCCGCCACTCAAAGCCACTGC-3'. The forward primer contains the NotI restriction site, Kozak initiation sequence, and the N-terminal of the hPRLr beginning with the ER leader sequence; and the reverse primer contains the EcoRI restriction site and the 3' end of the full-length long hPRLr. These primers render a fragment containing the hPRLr ER leader sequence, ECD and the TM domain. This fragment was isolated and inserted into pEF Tracer-V5/His (Invitrogen) using the NotI/EcoRI restriction sites. PCR amplification of all constructs were performed as previously described (43). Endogenous stop codons were removed to permit addition of the Myc or V5 tags. Clones were checked for amplification errors by dideoxynucleotide sequencing.
Cell Culture and Transient Transfections
CHO cells were maintained as previously described (57). HepG2 liver carcinoma cells, T47D human breast cancer cells and COS7 monkey kidney cells were maintained in DMEM with 10% FBS. Before stimulation with recombinant human PRL (obtained from the National Hormone and Peptide Program, National Institute of Diabetes and Digestive and Kidney Diseases), cells were rested in defined medium supplemented with 0.1% BSA (Sigma, St. Louis, MO). CHO cells were transfected following manufacturer’s instructions with FuGene (Roche Clinical Laboratories, Indianapolis, IN); COS7 and T47D cells were transfected following manufacturer’s instructions with Lipofectamine Plus (Invitrogen).
Immunoprecipitations and Immunoblot Analysis
To detect ligand-independent heterodimerization of long and intermediate hPRLrs, approximately 1 x 106 T47D and HepG2 cells in 100-mm plates were rested in defined media for 48 h before the addition of 0 or 150 ng/ml PRL. Cells were maintained on ice for up to 20 min to prevent internalization of receptor, lysed with 500 ml of Nonidet P-40 buffer, and centrifuged to remove debris. Lysates were immunoprecipitated with an anti-PRLr-Intermediate antibody (Zymed, South San Francisco, CA) that specifically recognizes the unique C-terminal region of the intermediate hPRLr at 4 C overnight. Parallel lysates were immunoprecipitated with normal rabbit IgG as a negative control. Immunoprecipitates were resolved by 10% SDS-PAGE, transferred to polyvinyldine difluoride (PVDF) membrane and immunoblotted with an anti-PRLr-ECD antibody that recognizes the ECD of the hPRLr. To detect total amount of hPRLr immunoprecipitated from T47D cells, parallel samples were treated as described above and immunoprecipitated with the anti-PRLr-ECD antibody at 4 C overnight and immunoblotted with as described above. To ensure that the PRLr was completely immunoprecipitated, cell lysates were immunoprecipitated with the anti-PRLr-ECD antibody at 4 C overnight and the supernatant was subjected to a second immunoprecipitation with the anti-PRLr-ECD antibody at 4 C overnight before immonoblotting as described above.
To detect ligand-independent hetero/homodimerization of hPRLr isoforms, COS7 cells were transiently transfected with hPRLr isoforms tagged with either Myc or V5 epitopes. Cells were treated as described above. Lysates were immunoprecipitated with an anti-Myc antibody (Upstate, Lake Placid, NY) at 4 C overnight. Immunoprecipitates were resolved by 10% SDS-PAGE, transferred to PVDF membrane and immunoblotted with an anti-V5 antibody (Invitrogen). Immunoblotting of whole-cell lysates verified expression of transfected proteins, and stripping and reprobing of blots verified expression and immunoprecipitation of Myc-tagged constructs.
To detect stability of the wild-type hPRLr in the presence of TM-ICD, and to detect phosphorylation of the TM-ICD construct, T47D cells were transiently transfected with vector or TM-ICD PRLr constructs. To detect stability of the wild-type hPRLr in the presence of TM-ICD, T47D cells were maintained in normal medium and maintained on ice for 20 min before cell lysis to prevent internalization of the receptor. These studies were also performed when T47D cells were rested for 48 h and then dosed for 24 h with 150 ng/ml PRL to ensure levels did not change under the same conditions used in the luciferase assays. To detect phosphorylation of TM-ICD, cells were arrested for 24 h and then treated with PRL for 20 min at 37 C. Cell lysates were immunoprecipitated with an anti-PRLr-ICD antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) corresponding to amino acids 323–622 of the hPRLr at 4 C overnight. Immunoprecipitates were resolved by 10% SDS-PAGE, transferred to PVDF membrane and immunoblotted with an antiphosphotyrosine antibody (Upstate) to detect phosphorylated tyrosine residues. Immunoblotting of whole-cell lysates verified expression of transfected proteins, and stripping and reprobing of blots verified immunoprecipitation of PRLr.
Detection and quantification of resolved proteins was conducted using a FUJI LAS-3000 imaging apparatus and MultiGuage software (FugiFilm, Valhalla, NY).
Reporter Gene Assays
The reporter construct LHRE-TK-Luc containing the Stat5 DNA-binding sites from the promoter region of ß-casein 5' to a luciferase reporter was a kind gift of R. J. M. Ross (Sheffield University, Sheffield, UK). CHO cells were plated at 70% confluency in 30-mm dishes and cotransfected using FuGene according to the manufacturer’s protocol with Renilla and LHRE-TK-Luc in addition to either 2 µg of vector, 1 µg of vector and 1 µg of wild-type hPRLr (wt), 1 µg of wt and 1 µg of ECD-TM 1 µg of wt and 1 µg of TM-ICD mutant PRLr. T47D cells were plated at 70% confluency in 30-mm dishes and cotransfected using Lipofectamine according to manufacturer’s protocol with Renilla and LHRE-TK-Luc in addition to 1 µg of vector, 1 µg of ECD-TM 1 µg of TM-ICD; or with Renilla and LHRE-TK-Luc in addition to 5 µg of vector or with increasing amounts of TM-ICD mutant PRLr (1, 2.5, or 5 µg) so that in each dish the amount of transfected vector or TM-ICD alone or in combination equaled 5 µg total. After transfection, cells were rested in defined media for 24 h and then stimulated with 0 or 100 nm dexamethasone and 150 ng/ml hPRL for 24 h. Dexamethasone treatment alone does not affect the LHRE-reporter activity, as previously reported (57). Luciferase assays were conducted by standard methods using the Dual Luciferase Assay System (Promega, Madison, WI) and the Luminoskan Ascent Type 392 (Thermo Labsystems, Inc., Franklin, MA). Data are reported as luciferase/Renilla ratio. Statistical analysis was generated by standard Student’s t test, and data are representative of one of three experiments conducted with triplicate transfections.
|
FOOTNOTES |
|---|
Disclosure: The authors have nothing to disclose.
First Published Online July 13, 2006
Abbreviations: CHO, Chinese hamster ovary; CypB, cyclophilin B; ECD, extracellular domain; EPO, erythropoietin; EPOr, EPO receptor; GHBP, GH binding protein; GHr, GH receptor; hPRLr, human PRLr; ICD, intracellular domain; PRL, prolactin; PRLr, PRL receptor; PRLBP, PRL binding protein; PVDF, polyvinyldine difluoride; TACE, TNF- converting enzyme; TM, transmembrane.
Received for publication March 8, 2006. Accepted for publication July 6, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
S1PRLr) containing only one fibronectin-like domain. Mol Endocrinol 16:2310–2322 converting enzyme (TACE) is a growth hormone binding protein (GHBP) sheddase: the metalloprotease TACE/ADAM-17 is critical for (PMA-induced) GH receptor proteolysis and GHBP generation. Endocrinology 141:4342–4348
-secretase cleavage. J Biol Chem 280:19331–19342This article has been cited by other articles:
 |
|
 |
|
  Y.-L. Xie, S. A. Hassan, A. M. Qazi, C. H. Tsai-Morris, and M. L. Dufau Intramolecular Disulfide Bonds of the Prolactin Receptor Short Form Are Required for Its Inhibitory Action on the Function of the Long Form of the Receptor Mol. Cell. Biol., May 15, 2009; 29(10): 2546 - 2555. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Fang, M. A. Rycyzyn, and C. V. Clevenger Role of c-Myb during Prolactin-Induced Signal Transducer and Activator of Transcription 5a Signaling in Breast Cancer Cells Endocrinology, April 1, 2009; 150(4): 1597 - 1606. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Shao, M. Nutu, B. Weijdegard, E. Egecioglu, J. Fernandez-Rodriguez, E. Tallet, V. Goffin, C. Ling, and H. Billig Differences in Prolactin Receptor (PRLR) in Mouse and Human Fallopian Tubes: Evidence for Multiple Regulatory Mechanisms Controlling PRLR Isoform Expression in Mice Biol Reprod, October 1, 2008; 79(4): 748 - 757. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Ben-Jonathan, C. R. LaPensee, and E. W. LaPensee What Can We Learn from Rodents about Prolactin in Humans? Endocr. Rev., February 1, 2008; 29(1): 1 - 41. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Kramer, W. Hanel, F. Shen, N. Isik, J. P. Malone, A. Maitra, W. Sigurdson, D. Swart, J. Tocker, T. Jin, et al. Cutting Edge: Identification of a Pre-Ligand Assembly Domain (PLAD) and Ligand Binding Site in the IL-17 Receptor J. Immunol., November 15, 2007; 179(10): 6379 - 6383. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-B. Jomain, E. Tallet, I. Broutin, S. Hoos, J. van Agthoven, A. Ducruix, P. A. Kelly, B. B. Kragelund, P. England, and V. Goffin Structural and Thermodynamic Bases for the Design of Pure Prolactin Receptor Antagonists: X-RAY STRUCTURE OF Del1-9-G129R-hPRL J. Biol. Chem., November 9, 2007; 282(45): 33118 - 33131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Stocco, C. Telleria, and G. Gibori The Molecular Control of Corpus Luteum Formation, Function, and Regression Endocr. Rev., February 1, 2007; 28(1): 117 - 149. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
