| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Gene Expression and Protein Biochemistry (M.A.I., C.A.S., S.H.K., S.-J.D., T.G.C., J.S., J.G.G., K.H.P) and Department of Cheminformatics (D.L.S., D.E.V.), Discovery Research, GlaxoSmithKline, Research Triangle Park, North Carolina 27709
Address all correspondence and requests for reprints to: Kenneth H. Pearce, GlaxoSmithKline, P.O. Box 13398, Five Moore Drive, Research Triangle Park, North Carolina 27709. E-mail: kenneth.h.pearce{at}gsk.com.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(ER) can be influenced by ligand binding. In turn, the conformational state of ER affects the ability of the receptor to interact with a wide variety of protein accessory factors. To globally investigate ligand-based cofactor recruitment activities of ER, we have applied a flow cytometric multiplexed binding assay to determine the simultaneous binding of ER to over 50 different peptides derived from both known cofactor proteins and random peptide phage display. Using over 400 ER-binding compounds, we have observed that the multiplexed in vitro peptide-binding profiles are distinct for a number of compounds and that these profiles can predict the effect that ER ligands have on various cellular activities. These cell-based activities include transcriptional regulation at an estrogen response element, MCF-7 cell proliferation, and Ishikawa endometrial cell stimulation. The majority of the compound-induced diversity in the peptide profiling assay is provided by the unique phage display peptides. Importantly, some of these peptides show a sequence relationship with the corepressor motif, suggesting that peptides identified via phage display might represent natural binding partners of ER. These in vitro:cellular correlations may in part explain tissue-specific activities of ER-modulating compounds. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Two different ER genes have been identified, yielding the two sequence-related protein products, ER and ERß (7, 8). Although ERß is an important participant in the estrogen response, several studies suggest that ER is responsible for much of the physiology associated with estrogens and estrogen replacement therapy (9, 10). It is commonly accepted that estrogens bind to the ligand-binding domain (LBD) of ER and elicit, among many other activities, an increase in transcriptional activity for genes containing some form of an estrogen response element (ERE) within the promoter region. This transactivation activity is enhanced by the ligand-induced association of cofactor or coactivator proteins, such as steroid receptor coactivator-1 (SRC-1) or cAMP response element binding protein-binding protein (CBP) (11).
Numerous crystal structures have confirmed that the most C-terminal helix, typically helix 12 known as the activation function or AF helix, of the receptor is prone to conformational flexibility (12, 13, 14). Agonist ligands such as estradiol and diethylstilbestrol (DES), tend to stabilize the AF helix into a position that exposes a hydrophobic patch or groove that can accept an amphipathic helix presented by cofactor proteins. These short helical regions, called NR boxes, usually occur in multiples within the cofactor protein and typically contain an LXXLL motif (where L is leucine and X is any amino acid) (15). Antagonist ligands, such as tamoxifen and raloxifene, force helix 12 into a position where it occupies the coactivator binding site and consequently blocks the intermolecular association of the coactivator amphipathic helix (12).
The conformational flexibility of helix 12 is an important feature of nuclear receptors and can allow for the design of ER-modulating ligands that shift the dynamic equilibrium between multiple states. In addition, nuclear magnetic resonance studies with another structurally related nuclear receptor have shown that ligands can affect the conformational dynamics in other regions of the LBD in addition to the AF helix (16). The conclusion of these structural studies was that subtle modifications of the ligand produce significant effects on the conformation of the LBD in multiple regions. In light of the emerging structural and functional work on ER and its LBD, there is an ongoing quest to find new selective ER modulators (SERMs) that may induce unique conformations. In addition to the well-known molecules raloxifene and tamoxifen, a new class of modulators might retain more of the beneficial effects of estrogen without the undesired side effects (17).
Similar to the conformational complexity of the LBD, ER performs a myriad of signal transduction functions once bound with ligand. The complexity of ER signaling is demonstrated by the observation that genes can be regulated either directly, through binding to EREs, or indirectly, through influencing the function of other proteins such as nuclear factor-
B (NF-B), activator protein-1 (AP-1), MAPK, or STAT5 (6, 18, 19, 20, 21). Concerning the well-studied ability of ER to repress NF-B-driven gene expression, the precise molecular mechanism for this activity remains unresolved and controversial (20). Several models have been proposed and one of these includes the direct physical interaction of ER with either the p50 or p65 subunit of NF-B (22, 23). In addition to interaction with transcription factors, it has been shown that activated ER can alter pathways in the cytoplasm, such as kinase phosphorylation cascades. These effects of ER are generally referred to as nongenomic activities. It is becoming more apparent that these functions are a significant part of the ER activity repertoire (24). The development of the next generation of SERMs will need to consider numerous ER activities.
To rapidly characterize the various protein-binding functions of ligand-activated ER as represented by peptide surrogates, we have used a method that can essentially reconstruct a simplified in vitro cellular environment for the receptor. This multiplexed technique allows simultaneous monitoring of up to 100 ligand-modulated interactions. Through the use of fluorescently distinct microsphere populations, individual peptides, for example peptides representing LXXLL NR boxes from coactivators or sequences identified via phage display, can be coupled to separate microsphere populations. Flow cytometric analysis of the microspheres simultaneously decodes each population and also detects receptor binding to respective peptides. Use of this fluorescent microsphere multiplexed binding assay has been previously demonstrated for nuclear receptor interactions (25), as well as several other applications (26, 27).
This binding-partner profile represents, in part, opportunities that the ER has to interact with partners in a cellular environment. In addition, phage display identified peptide sequences have the potential to represent currently unrecognized cellular interaction partners. These peptide interactions represent the starting signal for cellular outcomes driven through the ER, such as gene regulation, cytokine production, cell proliferation, and cell stimulation. To acquire more evidence that peptide binding profiles reflect ER biology, we have used a large number (>400) of ER-binding compounds that have varying degrees and directions of modulating activities and we have obtained both binding profiles and cell-based activities for these compounds. We have found that a number of peptide binding profiles correlate well with cellular activities. For this report, we have used several cell-based assays to explore the ability of ligand-activated ER to effect 1) transcription of a reporter gene via an ERE, 2) the proliferation of a breast cancer cell line, and 3) the stimulation of an endometrial-derived cell line.
As described in this paper, a simplified in vitro binding assay for monitoring the myriad of activities of the ER can be a great aid in helping to predict and understand the cellular outcomes of modulating ligands. Using the compounds as the link between these two types of assays, in vitro and cellular, we have been able to make distinct correlations between peptide binding events and cell-based assay activities imparted by novel ER ligands.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
are shown in Table 1
. Typically, with selection vs. ER LBD bound to an agonist ligand such as estradiol or DES, we readily identified LXXLL-containing peptides after a single round of sorting. However, not all peptides that bound to the ER LBD-estradiol complex contained the NR box motif. One example is the peptide denoted K7P2 as shown in Table 1. K7P2 binds to ER LBD in the presence of many different estradiol-like ligands with the same behavior as a classical LXXLL, but does not contain a typical leucine motif. In contrast, clones selected vs. antagonist-bound ER LBD, such as raloxifene, tamoxifen, idoxifene, and levermeloxifene, yielded only non-LXXLL peptide sequences. Several novel ER LBD compound-selective peptides were identified including I8P2 for idoxifene, R5P2 for raloxifene, and GW5P2 for GW5638 or its 4-hydroxyl metabolite, GW7604. Although we were able to successfully isolate peptides with the random 10-residue peptide library, it is evident from sibling DNA sequences of selected peptides that the size of the original library limits peptide diversity.
|
LBD-GW7604 complex, we made a second-generation phage display library on a phagemid vector for soft randomization or affinity maturation (32, 33, 34). This library was based on the sequence of the GW5P2 clone. Each of the ten residues was randomized with a bias toward the original codon using 80% of the bases from the original template sequence and an equal mixture of the three other bases at 20%. Three additional randomized codons were added to both ends of the original sequence. By sorting this library against ER LBD in the presence of GW7604 (or other compounds with an acrylate head group), we identified peptide sequences 47P3 and 33P4. Although this peptide library was re-randomized, the selected acrylate-sensing peptide sequences retained the original LF(E/D)RVWLRE(L/H) motif of GW5P2 (Table 1
). An additional leucine residue was selected in the three randomized codon regions at both the N and C terminus. These extra hydrophobic residues may account for the increased affinity of 47P3 for ER compared with the original GW5P2 peptide (demonstrated by increased binding signal; see Fig. 2B).
|
and Ref.25) were used. These peptides were typically about 25 amino acids in length with a biotin at the amino terminus.
Multiplexed-Peptide Binding Assay
To efficiently assess the effects of 405 ER-binding compounds on ER binding to 58 peptides representing the binding motifs of coactivator, coregulator, and corepressor proteins as well as compound-selective phage display peptide sequences, we employed a modified microsphere-based binding assay that has been described previously (25). All compounds within this set were determined to bind ER with a pKi 
6.0. We included several classic, well-characterized ER ligands such as estradiol, DES, raloxifene, 4-OH tamoxifen, idoxifene, levormeloxifene, lasofoxifene, and GW7604 (35) (Fig. 1) within the 405 compound set. Binding of Alexa532-labeled ER LBD is reported in mean fluorescence intensity units (MFI). Under the conditions of the assay, the MFI is proportional to affinity of peptide binding (25).
|
LBD binds to many of the peptides (Fig. 2A). Compounds may modulate the interaction of ER with peptide by enhancing or inhibiting this basal interaction. Typically, to demonstrate the effect of compound on ER peptide-recruitment activity, we subtract the basal value from the value obtained in the presence of compound. Therefore, if a compound enhances a binding interaction, the data point will have a value greater than zero and alternatively, if the interaction is inhibited it will have a value less than zero. Figure 2B highlights the binding profiles of ER LBD for 58 peptides in the presence of estradiol, DES, 4-OH tamoxifen, raloxifene, GW7604, idoxifene, lasofoxifene, and levormeloxifene. Each line represents binding data for one compound across the set of peptides. Other compounds examined are shown as gray lines in the background. Estradiol and DES enhanced the binding of ER LBD to coactivator and coregulator peptides as well as to the pan-agonist phage-identified peptide K7P2. In contrast, 4-OH tamoxifen, raloxifene, GW7604, idoxifene, and levormeloxifene all inhibited ER LBD basal binding to coactivator and coregulator peptides. 4-OH tamoxifen enhanced the binding of ER LBD to the tamoxifen-selective peptide sequence /ßV (29), and GW7604 promoted ER LBD binding to the peptides GW5P2, 47P3, and 33P4. Idoxifene, and to some degree levormeloxifene and lasofoxifene, enhanced ER LBD binding to I8P2.
Shape-Based Clustering of Peptide Binding Profiles
Many of the peptides fall into groups that either behave similarly in the presence of compound or do not bind ER LBD. To identify these peptide sets and reduce possible statistical bias, principal components analysis (PCA) was conducted using binding data from the 405 compound set. PCA reduced the number of peptides in the set from 58 to 10 representative peptides that account for approximately 99% of the variation in the data (see Data Reduction Methods). Figure 2C shows the same binding data from Fig. 2B after peptide selection by PCA. Interestingly, we have found that most of the coactivator binding trends for these compounds on ER LBD could be described by data from receptor-interacting protein 140 (RIP140) (699–723), RIP140 (922–946), transcriptional intermediary factor 2 (TIF2) (676–702), peroxisome proliferator-activated protein 
coactivator-1 (PGC-1) (130–154), and SHP (7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31). From the data shown in Fig. 2, B and C, it is evident that most of the compounds affect coactivator peptides in the same manner. In other words, if a compound promotes ER LBD binding to one LXXLL peptide it will most likely do so for other LXXLL peptides (especially for peptides that bound in the absence of compound). Subsequent clustering of the profile data was conducted, using the 10 peptide variables, as shown in Figs. 3 and 4.
|
|
, Fig. 4 shows the shape-based clustering results from the 405 binding profiles. A number of compounds yielded profiles that were relatively muted in terms of peptide recruitment. These compounds fall into cluster 1 (178 compounds) and cluster 2 (45 compounds). Clusters 3 (three compounds), 4 (14 compounds), and 5 (32 compounds) can be described as agonist or estradiol like. Compounds in these clusters show significant coactivator and K7P2 peptide recruitment with little effect on ER LBD binding to corepressor and most phage-derived peptides. Clusters 6 (two compounds), 7 (33 compounds), and 8 (59 compounds) are each unique and contain the tool compounds idoxifene, 4-OH tamoxifen, and raloxifene/lasofoxifene/levormeloxifene, respectively. Although ER LBD binding to coactivator peptides is inhibited for all the compounds that fall within clusters 6–8, the unique binding attribute of each of these clusters comes from the phage display peptides. Cluster 9 (1 compound) shows a profile somewhat similar to that for the 4-OH tamoxifen cluster 7. The final three clusters, cluster 10 (27 compounds), 11 (one compound), and 12 (10 compounds), contain variations of profiles that show binding to the GW5P2 peptide. In addition, many compounds in cluster 12 (and some found in cluster 10) also yield significant recruitment of 2AP3, which is a phage peptide identified with a GW7604-like ligand. Overall, this analysis shows that different ER ligands have the ability to differentially effect peptide recruitment and that these differences can be used to group compounds using a hierarchical clustering routine.
Correlation of Cellular Function with Peptide Binding Profile
Several cell-based assays were used to further characterize a subset of the 405 ER compounds. A transient transfection assay was used to assess the ability of compounds to drive reporter gene expression via the consensus vitellogenin ERE promoter (36). Compounds were evaluated for their ability to stimulate transcriptional activity (agonist mode, Fig. 5A). Those that did not show agonist activity were evaluated for their ability to antagonize or inhibit transcriptional activity of a subsaturating dose of estradiol (antagonist mode, Fig. 5B). To incorporate the pEC50 and pIC50 with efficacy values from the full dose-response curve fits, the potency measurement reported on the x-axis is the product of the two (where efficacy is a percentage relative to either estradiol or raloxifene standards). In cases where we have full dose-response curves, we chose to display the data in this manner because it incorporates both measured parameters. Trends are the same using only efficacy values, but the correlation with peptide binding data is improved when accounting for EC50 or IC50 value. As shown in Fig. 5A, a subset of ligands was examined for their effects on binding to the SRC-1(2) peptide (blue squares) and also for agonist activity at an ERE. These same ligands were also tested for binding to peptides from amplified in breast cancer 1 (AIB1) (607–631) (green triangles), K7P2 phage derived (red circles), PGC-1 (130–154) (yellow triangles), DAX1 (132–156) (cyan hexagons), and TIF2 (676–702) (red triangles). There is a general trend of increased ERE potency relative to coactivator peptide recruitment. Similarly, those compounds most effective at antagonizing the activity of estradiol show the most negative peptide recruitment values (Fig. 5B).
|
shows the potency of the compounds expressed as the product pEC50 or pIC50 (where pIC50 are negative values) and the normalized efficacy value; the y-axis shows peptide data for these same compounds where blue squares represent binding data for SRC-1(2) (676–700) peptide and other colored symbols represent five additional peptides. As shown, there is a general trend toward increased coactivator recruitment and increased ability to cause MCF-7 cell proliferation.
|
shows Ishikawa data for a subset of compounds plotted vs. SRC-1(2) peptide binding (blue squares) for the same set of compounds. Data for five other peptides are also shown (other colored symbols). Figure 7A shows that there is a general correlation between coactivator peptide recruitment and ability to stimulate Ishikawa cells. Figure 7B shows the relationship between Ishikawa stimulation and recruitment of three phage peptides, /ßV (tamoxifen-selective; pink triangles), I8P2 (idoxifene-selective; cyan circles), and GW5P2 (GW7604-selective; yellow squares). As shown, for compounds that showed significant recruitment of either the /ßV or I8P2 peptides, Ishikawa %Emax values were less than 20%. For compounds that recruit the GW5P2 peptide, the Ishikawa %Emax values were even lower (<5%).
|
, we have shown that there is a general correlation between coactivator/phage peptide binding and the modulation of ER-driven cell-based activities imparted by these compounds. To further explore this relationship, Fig. 8 displays cell-based measurements for ERE, MCF-7, and Ishikawa assays in the same clusters as defined by the peptide binding data represented in Fig. 4. Figure 8 shows that, in general, compounds that cluster together based on peptide profile also show similar cell-based activities. For several of the clusters, the cell-based data show highly consistent results within the cluster. For example, cluster 3 (with estradiol and DES) shows the expected high Ishikawa stimulation, MCF-7 proliferation and ERE activity. Cluster 7, containing 4-OH tamoxifen, shows low Ishikawa stimulation, MCF-7 proliferation and ERE activity. Compounds in cluster 10, which contains GW7604, show very low Ishikawa %Emax values (between 2 and 5%). Clusters 11 and 12, which like cluster 10 is defined by strong binding to the GW5P2-like peptides, also show a trend for low Ishikawa response. Compounds in cluster 8, which includes raloxifene, lasofoxifene, and levormeloxifene, yield relatively low Ishikawa %Emax values (between 5 and 20%) and show the ability to inhibit both MCF-7 proliferation and ERE activation.
|
Correlation of Cell-Based Compound Effects with Corepressor Binding
Several recent studies have examined the role of an ER:corepressor interaction in mediating the tissue-selective effects of tamoxifen and raloxifene (42, 43, 44). We used the 405 compound set to search for ligands that might enhance the interaction between ER LBD and a peptide derived from either nuclear hormone receptor corepressor (NCoR) or silencing mediator of retinoic acid and thyroid receptor (SMRT). As shown in Fig. 9, A and B, a number of compounds were shown to enhance the recruitment of both the NCoR 1 [amino acids (aa) 2040–2065] and SMRT 2 (aa 2329–2354) peptides. The compounds that were most successful at enhancing this interaction are those with an acrylate or carboxylate head group extension (like GW7604; shown as green squares). In addition, some novel head groups were also found to induce the same response and cluster with the acrylate-containing compounds (red squares). As shown in the two examples in Fig. 9, the same compounds that increased corepressor peptide binding also increased binding to the GW5P2 phage display peptide and the GW5P2-like peptide, 47P3.
|
:GW7604 complex revealed that these peptides share the same hydrophobic residue spacing as corepressor motifs (Fig. 10). This alignment shows that the typical corepressor motif of LXX(H/I)IXXX(L/I) is represented in the majority of the acrylate-sensing phage display peptides. In addition, a Leu at the –1 position and the basic amino acid at the +7 position were selected in both the primary random 10-residue peptide library and the second-generation GW5P2-based library. Selection from the latter library identified a Thr at the –2 position and a Leu at the –3 and +12 position to further extend the presumed amphipathic helix. Also, Phe and Trp were preferentially selected at the +1 and +5 positions, where the corepressor contains the smaller hydrophobic amino acids Leu and Ile, respectively.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
ligands has previously relied heavily on cell-based predictive assays. A number of studies, using protease digestion (45), crystallography (12, 46), and phage display (28, 29), have elegantly shed light on the molecular mechanism for ligand-induced receptor conformations for ER. Such studies have highlighted the importance of receptor dynamics in governing outcomes of ER-binding compounds. An important step in further understanding the details of ER signal transduction and the consequent discovery of new modulating ligands is further defining the link between receptor conformation and biological activity of SERMS. Here, we have used various ER-binding small molecules as tools to probe both receptor conformation, via peptide interactions, and cell-based activity. In general, these results show that peptide interaction profiles correlate well with, and could be used to predict, the biological consequence of ligand binding within a cell as described by a set of cell-based predictive assays. These correlations further highlight the dominant role that ligand-induced conformation and association with cofactor have on ER cellular function. There has been tremendous effort toward understanding the differential effects of SERMs and the effects that they cause at the cellular level (17). One approach has been to use differential gene expression to fingerprint and classify ER ligands based on their ability to alter RNA transcript levels (47, 48). It is becoming clear from these types of studies that gene expression patterns help explain why some ER-modulating ligands display both beneficial outcomes and undesired activities such as activity in uterine or breast tissue. Future work should allow assigning specific genes to particular biological effects elicited by SERMs, such as retention of bone mineral density (49). However, parsing through the enormous numbers of available and newly synthesized ER-binding compounds at the cellular level is a daunting task. The results presented here demonstrate that the ligand-induced conformation of ER can be probed through protein-peptide interactions and these binding profiles may be used to characterize and bin new SERMs to aid selection of novel compounds. Compounds that yield profiles unlike those of known SERMs can be rapidly identified and flagged for further cell-based or animal-based work. This relatively high-throughput technology should help guide ER and other nuclear receptor chemistries for fine-tuning modulators of the future.
Interestingly, we have made these in vitro to cellular correlations even though our multiplexed peptide binding assay uses purified LBD. In fact, our ER expression construct contains an abbreviated form of the LBD (aa 299–554) that deletes the extended strand after the AF helix (F domain). Although we observe excellent correlations, expanding efforts to include full-length receptor and peptide/protein/DNA binding partners that recognize other regions of ER (such as the AF-1/DNA-binding domain/hinge regions) should increase the amount of information garnered. It is well documented that ER has numerous activities driven through the N-terminal AF1 domain. For example, it has been shown that the AF1 domain is phosphorylated upon ligand binding and that other ligand-induced activities can occur via cytoplasmic signaling events (50). The relationships that we derive in this current study demonstrate that ligand-influenced AF2 function is a dominant feature of ER signal transduction, and it is likely that AF1 phosphorylation and its activities are linked to the AF2 in the context of the full-length receptor. Although our data suggest that most of the receptor function correlates with ligand-dependent AF2 function, it is important to consider the possibility that in vitro profiling may not completely correlate with ER LBD binding and function. A case could be made that nongenomic effects, possibly driven through an alternate signal transduction pathway such as recently shown for progestin (51), may yield discrepancies between LBD profiling and cell-based activities for ER ligands. Additionally, lack of correlation between peptide binding characteristics in vitro and the effect of compounds on a cell could be due to lack of cell penetration by a compound or due to metabolic alteration of a compound upon entering a cellular environment.
Although a number of studies have suggested differential coactivator recruitment by ER ligands (52, 53), we have found, under our experimental conditions, that binding characteristics for most cofactor-derived LXXLL peptides track with one another. Generally, all estradiol-like compounds proportionally increased LXXLL peptide binding relative to basal binding and all tamoxifen- or raloxifene-like compounds proportionally decreased LXXLL binding. Application of PCA deselected 39 of the 44 cofactor peptides used in this study. In other words, the five cofactor peptides shown in Fig. 2C were adequate to fully describe the compound-induced binding trends of all cofactor peptides. It is possible that very subtle differences in cofactor recruitment with various compounds may require full-length proteins, where sites distal to the coactivator cleft (on the receptor) and/or NR box (on the cofactor) may play a role. Clearly though, as noted by previous studies (54, 55), there are distinct LXXLL preferences for ER LBD where, as shown in Fig. 2, A and B, peptides from RIP140, SRC-1, TIF2, AIB1, PGC-1, SHP, and DAX yield the highest binding signal. It is interesting to note that four of these five peptides, in addition to the phage-derived peptide K7P2 (LLRWYL), show a weak sequence relationship of 
LKXLL, where represents a hydrophobic side-chain (Leu, Val, or Ile) and K represents Lys in addition to the other basic residues Arg, Gln, and His (Table 1) (56). The DAX peptides deviate slightly by containing a Tyr residue at the +2 position (ILYXM/LL). These peptide sequences, notably K7P2, DAX (66–90), and DAX (132–156), also suggest that ER is able to tolerate either a Met or Tyr in the +4 position. A full understanding of the detailed molecular recognition aspects of the preferential binding of certain peptides will require further structural and functional analysis.
Figures 5–7 show that there is a consistent relationship between compound-induced coactivator peptide recruitment and the effect of compound on ER-driven cell response. To more accurately show this correlation, we have assembled various cell-based activities for a subset of the tool compounds and organized these data into clusters based on similarity of peptide profiles (Fig. 8). As shown, most of the cell-based data are similar in each particular cluster. The most obvious exceptions are data shown in cluster 1. The peptide profiles for this cluster are relatively unchanged from the peptide profile of ER in the absence of compound (i.e. the change in relative fluorescent intensity is approximately zero for all peptides). Although the compounds in this cluster bind with relatively high affinity, they do not appear to induce a unique, stabilized conformation. These types of ligands may act as competitive antagonists by keeping the receptor in an apo-like state. In this situation, the ligand-bound LBD structure would be predominantly a dynamic state similar to the ligandless receptor. It will be interesting to pursue whether or not peptides or other binding partners will be able to further discriminate compounds that appear to induce a somewhat neutral conformation. With additional peptides, the correlation of peptide profile with cellular activity for ligands within this cluster may improve.
It has been noted that tamoxifen, although developed as an antiestrogen for the treatment of breast cancer, has estrogenic activities in certain tissues (57). In particular, tamoxifen exhibits minimal, but significant uterotrophic activity relative to raloxifene. With respect to the LXXLL cofactor peptides, both tamoxifen and raloxifene induce very similar peptide profiles (Fig. 2B). However, the binding profiles are clearly distinct with respect to the tamoxifen-selective peptide /ßV (29) (Fig. 2C). Interestingly, compounds that induce recruitment of the peptide also yield an Ishikawa stimulation value less than 20% (Fig. 7B). One explanation for this relationship may be that the tamoxifen-selective peptide represents an uncharacterized cofactor that may be expressed in uterine tissue but not breast tissue. Although it has been demonstrated by chromatin immunoprecipitation experiments (44), we were not able to detect a significant difference with SRC-1 peptide recruitment between tamoxifen and raloxifene in vitro. This may be related to a lack of an AF1 region on ER or lack of full-length SRC-1 cofactor as regions outside the LXXLL may play a partial role in recruitment. In addition, we were not able to detect a significant interaction between ER and corepressor peptide with either tamoxifen or raloxifene.
To even further discriminate compounds with low uterine stimulatory activity via peptide profiling, we have selected peptides specific for two compounds, idoxifene and GW5638, or its 4-hydroxyl version GW7604. These compounds yield Ishikawa stimulation values less than 15% relative to estradiol. As shown in Fig. 2, B and C, the two selected peptides, I8P2 and GW5P2, are specifically recruited to ER upon binding of idoxifene or GW5638/GW7604, respectively. In addition, our test compound set contains several molecules similar to idoxifene and GW5638/GW7604 and these compounds also induce recruitment of the selective peptides (Fig. 2C). Similar to the /ßV tamoxifen-selective peptide, comparison of I8P2 and GW5P2 peptide recruitment with Ishikawa cell stimulation, shows that compounds that induce ER binding to these peptides also exhibit uterine cell activity approximately less than 20% (Fig. 7B). From the comparison of peptide and Ishikawa activities, the GW5P2 peptide appears to predict an even lower uterine activity. Although only several compounds are shown, we have examined more than 15 compounds that induce recruitment of the GW5P2 peptide and all these compounds display an Ishikawa activity less than 5% relative to estradiol (data not shown).
Sequence alignment shows that there is a relationship between the set of phage peptides isolated against the ER:GW5638/GW7604 complex and the corepressor peptides, from both NCoR and SMRT (Fig. 10). Compound-induced recruitment shows that both these phage peptides and the corepressor peptides are recruited by same compounds in a linear relationship (Fig. 9, A and B). This suggests that these peptides may interact at the same site on ER and have the same mode of interaction with ER. It is possible that the GW5P2 and 47P3 peptides identified by phage display are affinity-optimized versions of corepressor peptides. If this is the case, the hydrophobic face of the amphipathic helix would be approximately one helical turn longer than the typical coactivator peptide and the interaction on the receptor would require repositioning of the AF helix as shown in the recent peroxisome proliferator-activated receptor/SMRT peptide crystal structure (58). As shown in Fig. 9, the binding signal for the GW5P2 and 47P3 phage peptides is approximately 3- to 5-fold higher than that for the corepressor peptides. One possibility for the increased signal (due to increased affinity) is that the selected Phe at +1, Val at +4, and Trp at +5 provide more hydrophobic surface area for packing against the extended corepressor cleft. It is interesting that the acrylate head group compounds optimally create this molecular recognition surface on ER. Using an LXXLL focused library, Connor et al. (59) also identified a similar peptide consensus from phage display results with ER:GW7604. Further structural and functional work will be required to elucidate the precise molecular recognition of these peptides and relevance of their similarity to the corepressor motif. From this example, and the fact that phage display peptides contribute a large degree of the prediction information for peptide profiling, it appears that the selected peptides might represent naturally occurring protein partners, some of which may remain to be discovered.
As mentioned briefly above, the compound-selective peptides offer a great deal of diversity to the peptide-binding profiles in this type of approach. For examining new ligands, the profiles can give clues that a new ligand is unique compared with tool compounds, even if a specific peptide for that new ligand is not in the peptide set. In fact, compounds that induce either novel or neutral profiles can be selected for further random peptide work to identify new peptides to add to the screening set. In contrast to a static picture provided by a crystal structure, this type of profiling analysis provides a sense of the activities of the receptor, which is a balance between various "on" and "off" states. In effect, the peptide interaction profile is sensitive to the dynamics of the receptor when bound with ligand.
Much of the detail for how ligand-induced receptor dynamics translates to signal transduction capacity in a cellular environment has yet to be completely defined. Our data add to the mounting evidence that the allosteric coregulator recruitment activity of ER, as well as other members of the nuclear receptor family, is a dominant feature of receptor function that ultimately links biological activity to the ligand. Visualization of the multitude of ligand-induced peptide profiles further emphasizes the structural and functional plasticity of the nuclear receptor LBD. Understanding how to manipulate the numerous pathway activities of ER with ligands, even with many current unknown factors, remains a goal for both drug and tool compound discovery. This microsphere-based compound profiling strategy, coupled with random peptide library technologies, can be a valuable tool to rapidly assess compound-induced conformations of ER and can aid in the early identification of unique potential drug candidates.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
was expressed in Escherichia coli, purified to homogeneity, and biotinylated similar to that described previously for ERß LBD (25). The expression construct consisted of aa 299–554 and mass spectrometry was used to verify the mass of the purified protein. Full-length ER was purchased from Invitrogen (Carlsbad, CA).
Selection of ER-Binding Compounds
All compounds were either synthesized in the GlaxoSmith-Kline Medicinal Chemistry Laboratories or were purchased from commercial sources. Solid compounds were dissolved in dimethylsulfoxide (10 mM) and subsequently diluted with buffer before use. An in vitro [3H]estradiol displacement assay was used to select ER-binding compounds from a large set of candidate compounds (data not shown). The compounds selected for further analysis in the microsphere-based peptide binding assay and the cell-based assays had pKi’s > 6. The average pKi for the 405 compounds in this test set was 7.2 with SD of 0.8. The compound set included several classic, well-characterized ER ligands such as estradiol, DES, raloxifene, 4-OH tamoxifen, idoxifene, levormeloxifene, lasofoxifene, and GW7604 (35) (Fig. 1).
Phage Display
A modified polyvalent phage display vector, fGWg3, was constructed using the original vector fTC (60). fGWg3 contains a flag tag epitope (AspTyrLysAspAspAspLys) in frame immediately upstream of the sequence for the M13 geneIII protein. This vector was also engineered to contain a dual BbsI site between the signal sequence of the M13 gene III and the flag tag. A random peptide library, fGWX10, was constructed by inserting a DNA cassette containing ten consecutive NNK codons, where N represents an equal molar ratio of A, C, G, and T and K represents G and T, into BbsI digested fGWg3. This library displays a random 10-residue peptide sequence with flanking regions as follows: NH2-EDGGSX10(GGGGS)3-gIII protein. The DNA cassette encoding the random peptide sequence was constructed using a synthetic ligase chain reaction procedure (61) with the following four oligonucleotides: 5'-TGGGTCCIII IIIIIIIIII IIIIIIIIII IIIIIIIGGA GGCGGGGGTA-3', 5'-GCGGCGGTGG AGGGTCTGGG GGAGGTGGCT CTGGGAAGAC CT-3', 5'-TTCCAGGTCT TCCCAGAGCC ACCTCCCCCA GACCC-3', and 5'-TCCACCGCCG CTACCCCCGC CTCCMNNMNN MNNMNNMNNM NNMNNMNNMN NMNNGGA-3', where I represents inosine, M represents equal molar of C and A, N represents equal molar of A, C, G, and T. E. coli MC1061 cell electroporation and phage library preparation have been described previously (62). After ligation and transformation, a small fraction of the library was plated on LB/tetracycline plates. The library contained approximately 1010 transformants and individual clones were analyzed by both DNA sequencing and phage ELISA using immobilized anti-Flag monoclonal antibody.
A second-generation phagemid display library was constructed based on the selected peptide sequence for ER plus GW5638 (Table 1). The library oligonucleotide was made to re-randomize the nucleotide sequence encoding the GW5638 selected peptide. An oligonucleotide for cassette mutagenesis was synthesized with base mixtures where each base position was fixed at 80% based on the original template sequence and an equal mixture of the three other bases was used at 20%. In addition, three NNS, where S represents an equal molar ratio of G and C, randomized codons were added to both ends of the re-randomized oligo segment. This oligonucleotide mixture sequence served as a template for PCR amplification using the following sequences as primers: 5'-ccg gcc atg gcc gaa gac ggt ggg-3' and 5'-tac cag gtt cca ggt ctt ccc aga gcc acc tcc ccc aga ccc tcc acc gcc gct acc ccc gcc tcc-3'. The PCR amplified cassette was ligated into a BbsI-digested, gel purified pGW BBS6g3 phagemid vector. The ligated library was electroporated into XL1-Blue cells and titered to yield 1.25 x 109 individual transformants. Random clones were selected and analyzed by DNA sequencing to show that 94% of the clones had correct inserts containing randomized DNA sequences.
In preparation for selection of specific peptide clones, 500 nM of either biotinylated ER LBD or biotinylated ER FL in the presence of 1 µM compound of interest was immobilized on a neutravidin-coated microtiter plate well. Protein was immobilized for at least 1 h at room temperature and then wells were washed three times with PBS. For the first round of selection, approximately 1012 phage particles were added in casein blocking buffer (0.89 g dibasic sodium phosphate, 2.5 g Hammarsten casein, and 2.9 g sodium chloride in 250 ml ddH2O) plus the compound of interest. Phage were allowed to bind for 2 h at room temperature. After washing with PBS/0.025% Tween 20 to remove unbound phage, bound phage were eluted with 0.1 N HCl plus 1 mg/ml BSA, the solution in the well was adjusted to pH 2.2 with glycine and then neutralized with 2 M Tris-HCl. An aliquot of the eluted and neutralized phage was added to 1 ml of confluent K91 cells. Cells were infected with phage particles for 15 min at 37 C, and then cultures were grown overnight at 37 C. The following day, phage were harvested by standard methods using 10% polyethylene glycol-8000/2.5 M NaCl in preparation for the next round of selection.
After each round of phage sorting, eluted phage were titered to determine population enrichment. Typically after two to three rounds of selection, enrichment was observed and individual clones were analyzed by DNA sequencing. Phage ELISA (63) was used to assess both peptide affinity for immobilized ER and the selectivity of the peptide for the receptor in the presence of ligand. Based on phage ELISA signal and selectivity vs. a panel of known SERMs, individual peptides were chosen for synthesis.
Synthetic Peptides
Biotin-containing peptide sequences derived from coactivator, corepressor, and coregulator proteins or identified by phage display were synthesized by the Synpep Corp. (Dublin, CA), or American Peptide Co. (Sunnyvale, CA). Each peptide was purified by reverse-phase HPLC. Amino acid composition and peptide concentration were determined by conventional methods for each peptide sequence. Typically, these peptides were biotinylated during synthesis at the 
-amino group of lysine of the G4SK C terminus. This orientation allows for attachment to a solid support similar to the direction displayed as a gIII fusion. Some of the peptides, such as R5P2 for raloxifene and 30P2 for GI5986, did not display high signal-to-background (at 10 nM ER LBD) as synthetic peptides relative to signal achieved with phage ELISA. It is possible that these peptides require some aspect of the gIII protein or some other phage coat protein as structural scaffolding to allow display of active peptide on phage, whereas the pure synthetic peptide is not active. Therefore, these weaker affinity synthetic peptides were not typically useful in the multiplexing format used in this report (see below). An alternative for future consideration is that the peptide binding site density on the surface of the microsphere can be tailored according to the affinity so that both high and lower affinity peptides can be used in the same multiplexed experiment (25).
Fluorescent Microspheres
Lumavidin-coated polystyrene microsphere populations (5.6 µm in diameter) with 106 biotin-binding sites per bead were purchased from the Luminex Corp. (Austin, TX).
Coupling of Biotinylated Peptides to Microspheres
Each biotinylated peptide was coupled to a specific population of microsphere as described previously (25). Briefly, in 96-well 1.2-µm filter-bottom plates (Millipore, Bedford, MA), Lumavidin-coated microspheres were washed in buffer (PBS containing 0.02% Tween 20, 0.1% BSA, 0.02% sodium azide and 1 mM dithiothreitol) and incubated with biotinylated peptide (2.0 µl of a 100 µM stock) for 30 min at room temperature in the dark. An additional 30-min incubation with free D-biotin (20 µl of 5 mM D-biotin) was conducted to stop the coupling reaction and block any unoccupied biotin-binding sites. The microspheres were washed twice, combined and resuspended in buffer. Quantitation of the coupling reaction was determined by flow cytometry as described previously (25). All peptide couplings were evaluated to assure that peptide density was at least 250,000 peptide molecules per microsphere. We have determined previously (25) that this peptide density minimizes the variation due to avidity or rebinding effects. The microsphere populations were coupled on the day of the binding experiment.
Coupling of ER LBD to Fluorochrome
The coupling of ER LBD to the reporter fluorochrome Alexa 532 (Molecular Probes, Eugene, OR) was performed essentially as described previously (25). Briefly, purified, biotinylated recombinant ER LBD was incubated with Strepavidin Alexa 532 at a 10:1 molar ratio of receptor to fluorochrome for 2.0 min. The coupling reaction was stopped by adding 20 µl free D-biotin (5 mM) and incubating for 30 min at room temperature in the dark.
Multiplexed ER LBD Binding Assay
All components (microspheres, ligand, Alexa 532-labeled ER LBD) were combined in the wells of a 96-well microtiter plate in a total volume of 75 µl. Each microtiter well contained 3,000 microspheres of a given population and up to 60 individual microsphere populations. The suspensions incubated for 1.5–2 h at room temperature in the dark. Fluorescence associated with each microsphere was directly measured by flow cytometry without washing.
Microsphere-associated fluorescence was measured using a LX-100 flow cytometer (Luminex Corp., Austin, TX) equipped with an XY platform for automated sampling from the wells of a 96-well microtiter plate. The LX-100 has two lasers. The red laser (633 nm) excites the fluorochromes embedded within the microsphere for population identification and the green laser (532 nm) excites the reporter fluorochrome on the nuclear receptor to indicate binding. Instrument calibration was conducted using LX-100 calibration microspheres purchased from the Luminex Corp. according to the manufacturer’s instructions. Microsphere fluorescence was determined per well from a minimum of 100 microspheres of each microsphere population and was gated to exclude doublets and aggregates using side scatter measurements. Microsphere populations were identified by their unique two-dimensional fluorescent profiles at 658 and 712 nm. ER LBD binding was measured at 575 nm and is represented in relative MFI units.
ERE Assay
The ERE assay was conducted by transiently transfecting a T47D breast cell line with a plasmid containing 1) the frog vitellogenin promoter which drives expression of a secreted placental alkaline phosphatase reporter gene, and 2) a plasmid that expresses the full-length ER gene. The vitellogenin promoter contains the consensus vitellogenin ERE sequence (F-vitERE) (36, 64). In this ERE assay, the presence of an ER agonist increases transcriptional activity of the reporter gene. Compounds were evaluated for their potential to either stimulate transcriptional activity (agonist mode, increasing concentrations of compound are incubated with cells in the absence of a competing compound) or inhibit transcriptional activity (antagonist mode, increasing concentrations of compound are incubated with cells in the presence of a fixed concentration of estradiol). Transcriptional activity was measured by acquisition of OD405 which indicates secreted placental alkaline phosphatase activity. An EC50 (agonist mode) or IC50 (antagonist mode) was determined for each compound tested. A compound’s ERE potency is defined as pEC50 x %max (agonist mode) or pIC50 x %min (antagonist mode).
MCF-7 Cell Proliferation Assay
Cell proliferation of MCF-7 breast cancer cells was determined by [3H]thymidine incorporation (37, 38). Compounds were tested for their potential to either stimulate MCF-7 proliferation or inhibit proliferation. An EC50 or IC50 was determined for each compound tested. A compound’s MCF-7 potency is defined as pEC50 x %cpmmax (relative to estradiol) or pIC50 x %cpmmin (relative to raloxifene) (cpm = radioactive counts per minute).
Ishikawa Endometrial Cell Stimulation Assay
The stimulation of human endometrial cells (Ishikawa cell line) was determined by increase in intrinsic alkaline phosphatase activity (39, 40). Compounds were tested for their potential to stimulate endometrial cells, as compared with estradiol. The %Emax is defined as 100 x (spectrophotometric reading at 405 nm with compound minus blank)/(spectrophotometric reading at 405 nm estradiol minus blank).
Data Reduction Methods
The multiplexed ER LBD binding assay yielded relative fluorescence intensities for 405 compounds and 58 peptides. Figure 2B shows these intensities (vertical axis) plotted for each peptide tested (horizontal axis). The points representing the intensities for any given compound were connected with line segments to yield a profile (Spotfire DecisionSite, Spotfire, Inc., Somerville, MA). This 58-dimensional data set was simplified by applying PCA to yield the peptides (dimensions) deemed to provide most of the covariance in the data set. PCA was performed using the JMP application (SAS Institute, Cary, NC), yielding eigenvectors representing the relative contributions of the principal components to the total covariance. The 10 eigenvectors with greatest magnitude, representing approximately 99% of the covariance were selected. From these, peptides with coefficients greater than 0.25 were chosen for profile analysis. This resulted in a reduced data set, yielding profiles having 10 peptide dimensions (Fig. 2C).
These peptide binding profiles were subjected to hierarchical clustering according to profile shape. For each profile of N responses, all N(N-1)/2 pairwise differences were calculated and placed into a list. The distance between two such lists could then be characterized by their Euclidian distance. Using such a metric, agglomerative clustering was applied (unweighted pairwise group mean average, Fig. 3). Drawing a boundary through the clustering tree at a level yielding 12 clusters, and plotting profiles in a trellis arrangement by cluster ID gave a useful classification of profiles (Fig. 4). Cell assay data were similarly plotted as profiles. These data were limited in availability, and so were plotted without line segments connecting the data points. By arranging cell assay data profiles in a trellis using the 12 cluster IDs, the correspondence of peptide profile shape to cell assay results can be visually analyzed (Fig. 8).
|
ACKNOWLEDGMENTS |
|---|
LBD protein used in this study, a number of chemistry teams at GlaxoSmithKline for synthesizing ER ligands, Lisa Miller for ligand-binding data, and Alan Payne and Greg Horesovsky for assay support and discussion. |
FOOTNOTES |
|---|
, estrogen receptor ; ERE, estrogen response element; LBD, ligand-binding domain; MFI, mean fluorescent intensity; NCoR, nuclear hormone receptor corepressor; NF-B, nuclear factor-B; PCA, principal components analysis; PGC-1, peroxisome proliferator-activated receptor coactivator-1 ; RIP140, receptor-interacting protein 140; SERMS, selective ER modulators; SHP, short heterodimer partner; SMRT, silencing mediator of retinoic acid and thyroid receptor; SRC-1, steroid receptor coactivator-1; STAT5, signal transducer and activator of transcription 5; TIF2, transcriptional intermediary factor 2. Received for publication November 5, 2003. Accepted for publication February 7, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and ERß, in estrogen target tissues in vivo through the use of an ER-selective ligand. Endocrinology 143:4172–4177, but not estrogen receptor ß, is involved in the regulation of the OPG/RANKL (osteoprotegerin/receptor activator of NF-B ligand) ratio and serum interleukin-6 in male mice. J Endocrinol 171:425–433[Abstract]
monitored by NMR spectroscopy: implications for nuclear receptor activation. J Mol Biol 298:187–194[CrossRef][Medline]
with phosphatidylinositol 3-OH kinase. Steroids 67:935–939[CrossRef][Medline]
B and steroid receptor-signaling pathways. Endocr Rev 20:435–459 B and the steroid hormone receptors: mechanisms of mutual antagonism. Mol Endocrinol 12:45–56B site. Nucleic Acids Res 25:2424–2429 and ER ß. Proc Natl Acad Sci USA 96:3999–4004
5 adrenal steroids. Endocrinology 127:2757–2762 provides evidence in support of a role for corepressors in regulating cellular responses to agonists and antagonists. Mol Endocrinol 16:1778–1792. Br J Cancer 87:449–456[CrossRef][Medline]
and ERß. J Biol Chem 275:35986–35993 and ß. Mol Cell Biol 19:8226–8239. Nature 415:813–817[Medline]
NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  A. Koppen, R. Houtman, D. Pijnenburg, E. H. Jeninga, R. Ruijtenbeek, and E. Kalkhoven Nuclear Receptor-Coregulator Interaction Profiling Identifies TRIP3 as a Novel Peroxisome Proliferator-activated Receptor {gamma} Cofactor Mol. Cell. Proteomics, October 1, 2009; 8(10): 2212 - 2226. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Joseph, B. M. Wittmann, M. A. Dwyer, H. Cui, D. A. Dye, D. P. McDonnell, and J. D. Norris Inhibition of prostate cancer cell growth by second-site androgen receptor antagonists PNAS, July 21, 2009; 106(29): 12178 - 12183. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J. Berrodin, K. C. N. Chang, B. S. Komm, L. P. Freedman, and S. Nagpal Differential Biochemical and Cellular Actions of Premarin Estrogens: Distinct Pharmacology of Bazedoxifene-Conjugated Estrogens Combination Mol. Endocrinol., January 1, 2009; 23(1): 74 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. H. Hong, H. Sun, C. H. Jin, M. Chapman, J. Hu, W. Chang, K. Burnett, J. Rosen, A. Negro-Vilar, and J. N. Miner Cell-Specific Activation of the Human Skeletal {alpha}-Actin by Androgens Endocrinology, March 1, 2008; 149(3): 1103 - 1112. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. G. Shearer, D. J. Steger, J. M. Way, T. B. Stanley, D. C. Lobe, D. A. Grillot, M. A. Iannone, M. A. Lazar, T. M. Willson, and A. N. Billin Identification and Characterization of a Selective Peroxisome Proliferator-Activated Receptor {beta}/{delta} (NR1C2) Antagonist Mol. Endocrinol., February 1, 2008; 22(2): 523 - 529. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Zwart, A. Griekspoor, M. Rondaij, D. Verwoerd, J. Neefjes, and R. Michalides Classification of anti-estrogens according to intramolecular FRET effects on phospho-mutants of estrogen receptor {alpha} Mol. Cancer Ther., May 1, 2007; 6(5): 1526 - 1533. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Heldring, T. Pawson, D. McDonnell, E. Treuter, J.-A. Gustafsson, and A. C. W. Pike Structural Insights into Corepressor Recognition by Antagonist-bound Estrogen Receptors J. Biol. Chem., April 6, 2007; 282(14): 10449 - 10455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Folkertsma, P. I. van Noort, A. de Heer, P. Carati, R. Brandt, A. Visser, G. Vriend, and J. de Vlieg The Use of in Vitro Peptide Binding Profiles and in Silico Ligand-Receptor Interaction Profiles to Describe Ligand-Induced Conformations of the Retinoid X Receptor {alpha} Ligand-Binding Domain Mol. Endocrinol., January 1, 2007; 21(1): 30 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Albers, B. Blume, T. Schlueter, M. B. Wright, I. Kober, C. Kremoser, U. Deuschle, and M. Koegl A Novel Principle for Partial Agonism of Liver X Receptor Ligands: COMPETITIVE RECRUITMENT OF ACTIVATORS AND REPRESSORS J. Biol. Chem., February 24, 2006; 281(8): 4920 - 4930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tamrazi, K. E. Carlson, A. L. Rodriguez, and J. A. Katzenellenbogen Coactivator Proteins as Determinants of Estrogen Receptor Structure and Function: Spectroscopic Evidence for a Novel Coactivator-Stabilized Receptor Conformation Mol. Endocrinol., June 1, 2005; 19(6): 1516 - 1528. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Ozers, K. M. Ervin, C. L. Steffen, J. A. Fronczak, C. S. Lebakken, K. A. Carnahan, R. G. Lowery, and T. J. Burke Analysis of Ligand-Dependent Recruitment of Coactivator Peptides to Estrogen Receptor Using Fluorescence Polarization Mol. Endocrinol., January 1, 2005; 19(1): 25 - 34. [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 |
