|
|
Proc Natl Acad Sci U S A. 2007 April 17; 104(16): 6846–6851. Published online 2007 April 11. doi: 10.1073/pnas.0605125104. | PMCID: PMC1871873 |
Copyright © 2007 by The National Academy of Sciences of the USA Chemistry, Pharmacology Discovery of a small molecule antagonist of the parathyroid hormone receptor by using an N-terminal parathyroid hormone peptide probe Percy H. Carter,* Rui-Qin Liu, William R. Foster, Joseph A. Tamasi, Andrew J. Tebben, Margaret Favata, Ada Staal, Mary Ellen Cvijic, Michele H. French, Vanessa Dell, Donald Apanovitch, Ming Lei, Qihong Zhao, Mark Cunningham, Carl P. Decicco, James M. Trzaskos, and Jean H. M. Feyen Pharmaceutical Research Institute, Bristol–Myers Squibb Company, Princeton, NJ 08543-4000 Received July 18, 2006. Once-daily s.c. administration of either human parathyroid hormone (PTH)-(1–84) or recombinant human PTH-(1–34) provides for dramatic increases in bone mass in women with postmenopausal osteoporosis. We initiated a program to discover orally bioavailable small molecule equivalents of these peptides. A traditional high-throughput screening approach using cAMP activation of the PTH/PTH-related peptide receptor (PPR) as a readout failed to provide any lead compounds. Accordingly, we designed a new screen for this receptor that used a modified N-terminal fragment of PTH as a probe for small molecule binding to the transmembrane region of the PPR, driven by the assumption that the pharmacological properties (agonist/antagonist) of compounds that bound to this putative signaling domain of the PPR could be altered by chemical modification. We developed DPC-AJ1951, a 14 amino acid peptide that acts as a potent agonist of the PPR, and characterized its activity in ex vivo and in vivo assays of bone resorption. In addition, we studied its ability to initiate gene transcription by using microarray technology. Together, these experiments indicated that the highly modified 14 amino acid peptide induces qualitatively similar biological responses to those produced by PTH-(1–34), albeit with lower potency relative to the parent peptide. Encouraged by these data, we performed a screen of a small compound collection by using DPC-AJ1951 as the ligand. These studies led to the identification of the benzoxazepinone SW106, a previously unrecognized small molecule antagonist for the PPR. The binding of SW106 to the PPR was rationalized by using a homology receptor model. Keywords: gene microarray, homology model, osteoporosis, bone Parathyroid hormone (PTH), a principal regulator of bone remodeling and calcium ion homeostasis, exerts its effects by binding and activating the PTH/PTH-related peptide receptor (PPR), a family B G protein-coupled receptor (GPCR). Notably, once-daily s.c. injection of either recombinant human PTH-(1–34) or PTH-(1–84) provides for increases in bone mineral density and decreased fracture risk in postmenopausal women. This observation has prompted broad interest in identifying orally bioavailable equivalents of PTH (reviewed in ref. 1). Herein, we disclose our initial efforts to find a small molecule ligand of the PPR. We began our efforts by screening for the ability to activate cAMP production in UMR106 cells, but this approach did not yield any tractable lead structures. Although this initial screen indicated that our compound collection did not contain any small molecule PPR agonists, it did not reveal any information about antagonist compounds. Extensive work with the family A GPCR small molecules, which typically bind to the juxtamembrane region of their receptors, has shown that trivial chemical modifications of antagonists can convert them into agonists ( 2). On the basis of this literature precedent, we assumed that a small molecule antagonist of the PPR could be a viable lead structure for the discovery of a small molecule agonist, provided that the antagonist bound to the juxtamembrane domain of the PPR. PTH-(1–34) is thought to bind by using a two-step sequence ( 1) in which the carboxyl-terminal domain of the peptide first binds to the extracellular domain of the PPR (, step A to B) and then the peptide N-terminal domain activates the receptor through interaction with its juxtamembrane domain (, step B to C). Thus, antagonists of a radiolabeled PTH-(1–34) might bind either to the N-terminal domain (and thereby block PTH binding) or to the juxtamembrane domain (and thereby block PTH function and perhaps allosterically modulate PTH binding), making PTH an unsuitable probe molecule for our purposes. We thus turned to examine the modified PTH-(1–14) derivatives first disclosed by Gardella and colleagues (leading refs. 3 and 4), which interact primarily with the juxtamembrane region of the PPR (, step A to D); we assumed that competitive antagonists of such peptides would themselves bind to the juxtamembrane region of the PPR. We set three short-term objectives for the research program: ( i) synthesize a truncated peptide that could serve as a viable tracer ligand in a binding assay, ( ii) document that the peptide was biologically equivalent to PTH-(1–34) in vitro and in vivo, and ( iii) validate that screening with this peptide could provide PPR antagonists. The successful results of our attempt to meet these three objectives are described in the sections that follow. Development of the Probe Peptide. Modification of the known PTH-(1–14) derivative 1 ( 3) such that it was suitable for use in a fluorescent spin polarization assay or standard radioactive binding assay provided peptides with nearly equivalent agonist potency ( 2– 4, Table 1). However, our attempts to use 2– 4 in the appropriately formatted binding assays were unsuccessful. We rationalized that this failure could be attributed to the suboptimal potency of the peptides, and so we examined the effects of substitution of aminoisobutryic acid (Aib) for Ala at positions 1 and 3, which has been reported by Gardella and colleagues ( 4) as potency-enhancing. As shown in the comparison of 4 and 5, this simple change, consisting of the addition of 28 Da to a 1,600-Da peptide, afforded a dramatic increase in agonist potency. | Table 1.Structure–activity relationships of peptides used in this study |
Encouraged by these results, we synthesized the three [Aib 1,3, Arg 11]-PTH derivatives 6– 8 ( Table 1), all of which were found to be potent agonists. Although we could not format a fluorescent spin polarization-based assay with 7, 8, or a more elaborate PTH-(1–21) analog (data not shown), we were able to use 125I- 6 in a standard radioactive binding assay in B28 cells [ supporting information (SI) Fig. 6], in which PTH-(1–34) and cold 6 afforded apparent IC 50 values of 0.69 ± 0.15 nM and 6.87 ± 0.90 nM, respectively. In saturation binding experiments, peptide 6 exhibited a Kd of 21.6 ± 5.2 nM and a Bmax of 603,000 receptors per cell. Importantly, 125I- 6 afforded the expected potencies when studied with our other peptide analogs in intact B28 cells ( Table 1), suggesting that the potency for cAMP formation of these peptides was likely related to their affinity for the peptide 6 binding site. The overlap of this binding site with the classical PTH binding site was shown both through displacement of 125I- 6 by PTH-(1–34) ( SI Fig. 6) and through the competitive antagonism of 6 by PTH-(3–34) in the cAMP assay ( SI Fig. 7A). Since the completion of our work, detailed pharmacological studies from other laboratories ( 5, 6) have shown that 14-mer and 15-mer peptides related to DPC-AJ1951 interact predominantly with the juxtamembrane and not the N-terminal domain of the PPR. These studies further strengthen the hypothesis that DPC-AJ1951 can serve as a probe of the juxtamembrane domain. Biological Characterization of the Probe Peptide. Because peptide 6 (hereafter referred to as DPC-AJ1951) met our basic criteria as a probe of the juxtamembrane domain of the PPR, we moved to study it further. As shown above ( Table 1), DPC-AJ1951 was a full agonist of cAMP production in cell lines containing either human or rat PPR. In addition, DPC-AJ1951 induced intracellular Ca 2+ mobilization in HEK 293 cells transfected with the human PPR (EC 50 26 ± 14 nM) (see SI Fig. 7B). As expected, DPC-AJ1951 did not elicit responses from cells that did not contain the PPR. Likewise, DPC-AJ1951 did not show any significant binding to the following selection of family A and B GPCRs when studied at concentrations up to 10 μM: calcitonin receptor, glucagon-like peptide 1 (GLP-1) receptor, adrenergic α1A, adrenergic α1B, histamine H1, dopamine D2, and 5HT2A. To broaden the comparison of signaling induced by DPC-AJ1951 and PTH-(1–34), we turned to microarray analysis. When 100 nM PTH-(1–34) was administered to UMR106 cells under low serum conditions, gene expression changes were time-dependent, with global expression changes increasing over time with 279, 416, and 514 reporters of 8,799 seen changing at P < 0.01 at 2, 8, and 24 h, respectively; likewise, 100 nM DPC-AJ1951 induced time-dependent changes, with 264, 472, and 479 reporters of 8,799 changing at P < 0.01 at 2, 8, and 24 h, respectively [see SI Table 3 for a listing of the significant transcriptional changes elicited by PTH-(1–34) and DPC-AJ1951]. The microarray results with DPC-AJ1951 were confirmed with real-time PCR for a selection of seven known markers of PTH action by using RNA from an experiment performed subsequent to that analyzed with microarrays, with UMR106 cells grown freshly from frozen stock ( Table 2). Notably, the transcriptional activity of DPC-AJ1951 was modulated by the PDE4b inhibitor rolipram ( SI Fig. 8), consistent with the inhibition of a negative feedback pathway. | Table 2.Confirmation of microarray results with real-time PCR |
The patterns of transcriptional activity elicited by probe peptide DPC-AJ1951 and PTH-(1–34) after 2 h were highly similar (): The responses showed a 98% correlation for transcripts having a P value of <0.01 in both treatments. The excellent correlation was maintained after 8 h of treatment, but the difference in potency between the peptides became evident by 24 h ( SI Fig. 9). Significantly, the correlation between the responses to 14-mer DPC-AJ1951 and PTH-(1–34) was higher than between the responses to partial agonist PTH-(3–34) and PTH-(1–34) ( SI Fig. 10 and Table 4). The transcriptional regulation induced by both DPC-AJ1951 and PTH-(1–34) was consistent with that reported for PTH-(1–34) after our work was completed ( 7) ( SI Table 4). The microarray data speak both to the activity and the selectivity of the 14-mer peptide: Not only does DPC-AJ1951 induce almost all of the changes induced by PTH-(1–34), it does not induce any substantial changes not induced by PTH-(1–34) (see SI Fig. 11 and Table 3), consistent with our initial in vitro GPCR counterscreening. Thus, even though the sequence of DPC-AJ1951, (Aib 1,3, Nle 8, Gln 10, Arg 11, Ala 12, Tyr 14)-PTH-(1–14)NH 2, is only 50% identical to that of native PTH-(1–14)NH 2, it is >100,000-fold more potent than PTH-(1–14) in stimulating cAMP formation in SaOS-2 cells and appears to have equivalent selectivity to the parent peptide, PTH-(1–34). To extend the in vitro findings described above, we studied 14-mer peptide DPC-AJ1951 in classical ex vivo models of PTH action. DPC-AJ1951 stimulated osteoclast-mediated bone resorption, as measured by the release of 45Ca from fetal rat long-bone explant cultures prelabeled with 45Ca in utero ( A). Likewise, in neonatal mouse parietal bone explants, DPC-AJ1951 caused a concentration-dependent decrease in collagen synthesis (incorporation of [ 3H]proline, B) and stimulation of cell proliferation (increase in [ 3H]thymidine, C). These data substantially extend the earlier report by Shimizu et al. ( 4) that a different Aib 1,3-PTH-(1–14) fragment can inhibit bone mineralization in embryonic mouse metatarsals in vitro. Although DPC-AJ1951 was perfectly stable in buffer or saline solution, it exhibited a relatively short half-life in rat plasma (≈20 min) or rat whole blood (≈15 min) when incubated at 37°C. Accordingly, we focused our attention on an acute in vivo model of PTH action in the rat. The peptide DPC-AJ1951 normalized serum calcium levels in a dose- and time-dependent fashion when administered to thyroidparathyroidectomized rats by continuous s.c. infusion ( D). The 14-mer DPC-AJ1951 was ≈40-fold less potent (molar basis) than PTH-(1–34) in this in vivo model of bone resorption (cf. D and SI Fig. 12), which is consistent with its ≈10-fold weaker potency in the aforementioned in vitro and ex vivo assays of PTH action ( Table 1 and A–C). This result demonstrates in vivo activity for peptide in the PTH-(1–14) class. Taken together, the results from our biological characterization of DPC-AJ1951 suggest that the extensive structural modifications embodied in DPC-AJ1951 enable this shortened peptide to bind and fully activate the PPR in a selective manner. Thus, although the overall potency and bioavailability of 14-mer DPC-AJ1951 are not equivalent to those of PTH-(1–34), † it appears to exhibit the same complement of pharmacological activity ( and ). Accordingly, we consider DPC-AJ1951 to be a validated probe of the agonist-binding site on the PPR. Use of the Probe Peptide in Small Molecule Screening. Having confirmed through several different mechanisms that DPC-AJ1951 induced qualitatively similar biological responses to the clinically efficacious PTH-(1–34), we used it to screen a compound collection consisting of structurally diverse chemotypes known to inhibit multiple protein target classes (GPCRs, proteases, kinases, etc). Herein we describe the initial characterization of the most potent compound identified in this initial screen, r-5-(2- E-cyclopropylvinyl)- t-3-ethyl-6,7-difluoro-5-(trifluoromethyl)benzo[e][1,4]-oxazepin-2( 1H, 3H, 5H)-one (hereafter SW106). Compound SW106 displaced 125I-DPC-AJ1951 with an IC 50 of 0.99 μM in B28 cells ( A). When applied to B28 cells or SaOS-2 cells alone, SW106 did not induce a cAMP response, suggesting that it was an antagonist. Consistent with this assumption, SW106 antagonized the cAMP response induced by DPC-AJ1951 in several cell lines, including those containing human and rodent variants of the PPR (see B for B28 cells; see SI Fig. 13 for others). SW106 was also found to be an antagonist of PTH-(1–34), but it exhibited ≈10-fold weaker potency against the longer peptide ( B). Whereas SW106 is not as potent as DPC-AJ1951 in the binding assay, it is more potent than several of the peptides listed in Table 1, despite its smaller size (molecular mass of 343 Da for SW106 vs. ≈1,600 Da for a typical PTH 14-mer), and is only ≈10-fold weaker than the most optimized PTH-(1–14) antagonist reported in the literature ( 8). We also studied the mode of antagonism by SW106. In a Schild-type analysis, SW106 weakened the apparent cAMP EC 50 of DPC-AJ1951 without changing the cAMP E max ( C), suggesting that it was a competitive antagonist. Consistent with this proposed competitive mechanism, the addition of 5 μM SW106 weakened the apparent affinity of DPC-AJ1951 without altering the total bound: the IC 50 values of DPC-AJ1951 for displacing 125I-DPC-AJ1951 were 6.9 ± 0.9 nM and 55.4 ± 9.2 nM in the presence of 0 μM and 5 μM of SW106, respectively ( SI Fig. 14). Given the promiscuity of many small molecule GPCR antagonists, we examined the selectivity profile of SW106. When studied at concentrations of 10 μM and above, we found that SW106 neither activated the human GLP-1 receptor nor blocked the actions of GLP-1 or nonnative peptide ligands (including truncated variants). Given that the GLP-1 receptor is a family B GPCR that couples through GαS, these data suggest that the ability of SW106 to block PTH-initiated cAMP accumulation is specific to its actions on the PPR and not related to its ability to interfere with other signaling mechanisms. We also studied the binding of SW106 to a panel of family A GPCRs. When tested at concentrations of 20 μM, the compound did not exhibit any activity against a number of these receptors (CC chemokine receptors 1, 2, 3, 4, and 5; CXC chemokine receptors 2 and 3; adrenergic α1B and α2A; dopamine D2; histamine H1; muscarinic M5; and serotonin 5HT2C), but did exhibit weak binding to serotonin 5HT2A (IC50 = 15 μM), adrenergic α1D (IC50 = 13.8 μM), neurokinin-2 (IC50 ≈ 28 μM), and adrenergic α2C (IC50 = 10.4 μM). Thus, SW106 displays higher affinity for the PPR relative to the other GPCRs studied. The lead compound SW106 embodies an unusual 4,1-benzoxazepinone ring structure as a constrained dipeptide motif and has been described as a member of an orally bioavailable series of HIV reverse transcriptase inhibitors ( 9). After identifying it as a PPR antagonist, we screened ≈60 additional representatives of this specific compound class to establish the initial structure-activity relationships in this series (data not shown). These data confirmed that multiple members of the structural class had activity against the PPR and that the chemotype was relatively tolerant to certain changes (e.g., halogenation pattern on aromatic ring, terminal appendage on olefin, alkyne exchange for olefin). Moreover, this screen identified the ethyl substituent as a key determinant of PPR binding: Although it could be altered to other small alkyl groups (methyl, n-propyl, isopropyl, cyclopropyl), alteration to phenyl (large, aromatic) or hydrogen (small) reduced activity dramatically. In addition, the relative stereochemical projection of the ethyl relative to trifluoromethyl was important, in that the cis-relationship allowed for maintenance of PPR activity, whereas the trans-relationship ablated that activity. To understand better the actions of SW106 at the PPR, we constructed a new homology model of the PPR by using the x-ray structure of bovine rhodopsin [Protein Databank (PDB) ID code 1F88] as a template ( SI Fig. 15). The PPR antagonist SW106 was manually docked into a hydrophobic/aromatic pocket between transmembrane domains 3, 4, 5, and 6 of the modeled PPR (; see also SI Fig. 16). The illustrated binding pose is consistent with the structure-activity relationships of SW106 ( vide supra); it is also consistent with the aforementioned selectivity of SW106 for the PPR over the GLP-1 receptor, as five of eight key contact residues vary in the illustrated pocket between the two receptors (see SI Fig. 17 for the sequence alignment). The hypothesis that SW106 might bind to residues on TM6 (specifically Met-425 and Tyr-429) allows us to rationalize other experimental observations. First, because Met-425 has been identified in photochemical cross-linking and site-directed mutagenesis experiments as a contact site for the first two N-terminal residues of PTH and/or PTH-related protein ( 10– 12), it suggests that SW106 blocks the binding of the N terminus of PTH; this is consistent with the ability of SW106 to antagonize the function of the C-terminally truncated series PTH-(1–34), PTH-(1–14), and PTH-(1– 11) (data with the 11-mer not shown). Second, it is in accord with the ability of SW106 to modify signaling by the PPR, as TM6 has been shown to be a key mediator of PPR activation ( 13, 14). The hypothesis that SW106 binds to a site contacted by Val-2 of PTH, the modification of which is known to enable switching between agonism and inverse agonism/antagonism ( 8, 10, 12), suggests that chemical modification of SW106 may enable this structure to activate the PPR rather than blocking its activation. Indeed, several independent lines of research with family A GPCRs have documented that this antagonist/agonist “pharmacological switching” can be achieved through sometimes surprisingly minor chemical modifications ( 2). Further exploration of the structure-activity relationships and pharmacological behavior of the SW106 chemical class is warranted: Given that the PPR plays a key role in a number of human disease states, optimization of SW106 to a potent, orally bioavailable agonist, antagonist, or inverse agonist could provide a compound of potential clinical utility. Peptide Synthesis. The reference peptide 1 and previously unrecognized analogs 2–8 were synthesized on an automated peptide synthesizer, globally deprotected, purified by reverse-phase HPLC, and characterized by electrospray MS and 1H-NMR. DPC-AJ1951 was iodinated to a specific activity of 2,000 Ci/mmol (1 Ci = 37 GBq) by using the lactoperoxidase method (Amersham Biosciences, Piscataway, NJ). Cell Culture. Cells were cultured at 37°C in T-75 flasks in DMEM supplemented with 10% FBS. Cells were treated with fresh media for 12–24 h before assay. cAMP Stimulation. Culture flasks containing nonconfluent SaOS-2 (or UMR106, B28, CLO153, or COS-7 with transiently transfected PPR) were rinsed with two or three washes of warm (37°C) PBS. Prewarmed EDTA (10–15 ml, 37°C) was added, and the flask was incubated (37°C, 5–10 min). The cells were counted while being centrifuged at 300 × g for 5 min and then resuspended in stimulation buffer (Hanks' balanced salt solution/5 mM Hepes Cellgro/0.1% BSA/0.5 mM 3-isobutyl-1-methylxanthine, pH 7.4). The AlphaScreen used to detect cAMP was performed as described by the manufacturer (Packard, Groningen, The Netherlands). Competition Binding. HKRK-B28 cells were harvested at 80–95% confluency by trypsinization and then plated at 30,000 cells per well in 96-well plates in 100 μl of media (high-glucose DMEM and 10% FBS supplemented with Hepes and Pen/Strep). Plates were incubated overnight at 37°C and 5% CO2. Compound (3 μl of DMSO solution, 50X) was added to 110 μl of binding buffer (50 mM Tris/100 mM NaCl/2 mM CaCl2/5 mM KCl/5.5% FBS, pH 7.7) and mixed. Media was discarded, and 75 μl of compound in binding buffer was added per well. 125I-6 was prepared in binding buffer so that ≈100,000 cpm (100 pM) was added per well in 25 μl. The plates were incubated for 3 h at room temperature before being harvested with three washes of 100-μl binding buffer. Fifty microliters of Microscint20 (Packard) was added per well, and the plates were covered with TopSeal (Packard). The plates were counted after ≈1 h on a TopCount (Packard) instrument. Microarray Cell Culture and Treatment. UMR106 cells were grown in DMEM and 10% FBS to confluency and switched to medium with 0.1% FBS 24 h before treatment. Solutions of compound, PTH-(1–34), and peptide were added to the media to achieve the concentrations indicated in the results. Expression Profiling. Total RNA isolations were performed by using the RNeasy (Qiagen, Valencia, CA) purification system according to the manufacturer's instructions. Expression profiling of RNA samples was performed by using the Affymetrix (Santa Clara, CA) U34A array following standard protocols. Image acquisition used GeneChip 5.0 (Affymetrix). Resulting CEL-file outputs were imported into Resolver 4.0 (Rosetta Biosoftware, Seattle, WA) for analysis. Real-Time PCR. Template cDNA was generated by using the Advantage RT-PCR kit according to the manufacturer's (Clontech, Mountain View, CA) instructions by using random hexamers and 1 μg of DNaseI-treated total RNA. Taqman-based RT-PCR was performed with a 7900HT according to the manufacturer's (Applied Biosystems) instructions with relative expression levels determined by serial dilution. In Vivo and ex Vivo Assays. Ex vivo bone resorption and formation assays were performed as described by us in ref. 15. Thyroparathyroidectomy surgery was performed on male Sprague–Dawley rats (200 g). Animals were supplemented with thyroxin (4 μg per rat, s.c.) three times a week. Animals with serum calcium levels between 5 and 8 mg/dl were used. One week after surgery, Alzet (Palo Alto, CA) minipumps were implanted s.c. for delivery of peptides at the concentrations indicated (see D legend). Serum samples were analyzed for total calcium by using the Calcium System Pack kit (Roche Diagnostics, Indianapolis, IN). We thank Prof. Thomas Gardella (Harvard University, Boston, MA) for providing the B28 cells; Dr. Nilsa Graciani for the resynthesis of DPC-AJ1951; Dr. Dean Wacker for the resynthesis of SW106; Dr. Pamela A. Benfield and D. Ellis for the cell culture used to obtain RNA; J. Bunville for design of primers; and Jing Chen, Melissa Yarde, Celeste Tawmley, and Ding Ren Shen for GPCR selectivity screening. | | Aib | aminoisobutryic acid | | | GLP-1 | glucagon-like peptide 1 | | | GPCR | G protein-coupled receptor | | | PPR | PTH-related peptide receptor | | | PTH | parathyroid hormone. |
1. Carter PH, Schipani E. Endocr Metab Immune Disord Drug Targets. 2006;6:59–76. [PubMed] 2. Beeley NR. Drug Discov Today. 2000;5:354–363. [PubMed] 3. Shimizu M, Carter PH, Khatri A, Potts JT, Jr, Gardella TJ. Endocrinology. 2001;142:3068–3074. [PubMed] 4. Shimizu N, Guo J, Gardella TJ. J Biol Chem. 2001;276:49003–49012. [PubMed] 5. Dean T, Linglart A, Mahon MJ, Bastepe MJ, Jüppner H, Potts JT, Jr, Gardella TJ. Mol Endocrinol. 2006;20:931–943. [PubMed] 6. Castro M, Nikolaev VO, Palm D, Lohse MJ, Vilardaga JP. Proc Natl Acad Sci USA. 2005;102:16084–16089. [PubMed] 7. Qin L, Qiu P, Wang L, Li X, Swarthout JT, Soteropoulos P, Tolias P, Partridge NC. J Biol Chem. 2003;278:19723–19731. [PubMed] 8. Shimizu N, Dean T, Tsang JC, Khatri A, Potts JT, Jr, Gardella TJ. J Biol Chem. 2005;280:1797–1807. [PubMed] 9. Cocuzza AJ, Chidester DR, Cordova BC, Klabe RM, Jeffrey S, Diamond S, Weigelt CA, Ko SS, Bacheler LT, Erickson-Viitanen SK, Rodgers JD. Bioorg Med Chem Lett. 2001;11:1389–1392. [PubMed] 10. Behar V, Bisello A, Bitan G, Rosenblatt M, Chorev M. J Biol Chem. 2000;275:9–17. [PubMed] 11. Bisello A, Adams AE, Mierke DF, Pellegrini M, Rosenblatt M, Suva LJ, Chorev M. J Biol Chem. 1998;273:22498–22505. [PubMed] 12. Gensure RC, Carter PH, Petroni BD, Jüppner H, Gardella TJ. J Biol Chem. 2001;276:42692–42699. [PubMed] 13. Vilardaga JP, Lin I, Nissenson RA. Mol Endocrinol. 2001;15:1186–1199. [PubMed] 14. Sheikh SP, Vilardaga JP, Baranski TJ, Lichtarge O, Iiri T, Meng EC, Nissenson RA, Bourne HR. J Biol Chem. 1999;274:17033–17041. [PubMed] 15. Staal A, Frith JC, French MH, Swartz J-A, Gungor T, Harrity TW, Tamasi J, Rogers MJ, Feyen JH. J Bone Miner Res. 2003;18:88–96. [PubMed]
|
This article has been 
