Structural Basis for Inhibition of Protein-tyrosine Phosphatase 1B by Isothiazolidinone Heterocyclic Phosphonate Mimetics
2006; Elsevier BV; Volume: 281; Issue: 43 Linguagem: Inglês
10.1074/jbc.m606873200
ISSN1083-351X
AutoresJ. Paul, Lucie Gonneville, Milton C. Hillman, Mary Becker-Pasha, Min Wei, Brian G. Reid, Ronald M. Klabe, Eddy W. Yue, Brian Wayland, Brent Douty, Padmaja Polam, Zelda R. Wasserman, Michael J. Bower, Andrew P. Combs, Timothy C. Burn, Gregory Hollis, Richard Wynn,
Tópico(s)Biochemical and Molecular Research
ResumoCrystal structures of protein-tyrosine phosphatase 1B in complex with compounds bearing a novel isothiazolidinone (IZD) heterocyclic phosphonate mimetic reveal that the heterocycle is highly complementary to the catalytic pocket of the protein. The heterocycle participates in an extensive network of hydrogen bonds with the backbone of the phosphate-binding loop, Phe182 of the flap, and the side chain of Arg221. When substituted with a phenol, the small inhibitor induces the closed conformation of the protein and displaces all waters in the catalytic pocket. Saturated IZD-containing peptides are more potent inhibitors than unsaturated analogs because the IZD heterocycle and phenyl ring directly attached to it bind in a nearly orthogonal orientation with respect to each other, a conformation that is close to the energy minimum of the saturated IZD-phenyl moiety. These results explain why the heterocycle is a potent phosphonate mimetic and an ideal starting point for designing small nonpeptidic inhibitors. Crystal structures of protein-tyrosine phosphatase 1B in complex with compounds bearing a novel isothiazolidinone (IZD) heterocyclic phosphonate mimetic reveal that the heterocycle is highly complementary to the catalytic pocket of the protein. The heterocycle participates in an extensive network of hydrogen bonds with the backbone of the phosphate-binding loop, Phe182 of the flap, and the side chain of Arg221. When substituted with a phenol, the small inhibitor induces the closed conformation of the protein and displaces all waters in the catalytic pocket. Saturated IZD-containing peptides are more potent inhibitors than unsaturated analogs because the IZD heterocycle and phenyl ring directly attached to it bind in a nearly orthogonal orientation with respect to each other, a conformation that is close to the energy minimum of the saturated IZD-phenyl moiety. These results explain why the heterocycle is a potent phosphonate mimetic and an ideal starting point for designing small nonpeptidic inhibitors. Protein-tyrosine phosphatase 1B (PTP1B) 2The abbreviations used are: PTP1B, protein-tyrosine phosphatase 1B; Tyr(P), phosphotyrosine; IR, insulin receptor; OTCA, 2-(oxalyl-amino)-thiophene-3-carboxylic acid; CMBA, 2-carboxymethoxybenzoic acid; DFMP, difluoromethylphosphonic acid; IZD, isothiazolidinone; TCEP, tris(2-carboxyethyl)phosphine hydrochloride; PDB, Protein Data Bank; r.m.s.d., root mean square deviation. is considered to be one of the best validated drug targets for the treatment of type II diabetes. The enzyme is localized to the cytoplasmic face of the endoplasmic reticulum where it negatively regulates insulin signaling by dephosphorylating phosphotyrosine (Tyr(P)) residues in the kinase regulatory domain of the insulin receptor (IR) (1Frangioni J.V. Beahm P.H. Shifrin V. Jost C.A. Neel B.G. Cell. 1992; 68: 545-560Abstract Full Text PDF PubMed Scopus (508) Google Scholar, 2Saltiel A.R. Kahn R. Nature. 2001; 414: 799-806Crossref PubMed Scopus (3972) Google Scholar). Mice lacking the homolog of PTP1B have lower blood glucose levels and improved insulin responsiveness compared with normal and diabetic mice through enhanced IR signaling in peripheral tissues (3Elchebly M. Payette P. Michaliszyn E. Cromlish W. Collins S. Loy A.L. Normandin D. Cheng A. Himms-Hagen J. Chan C.-C. Ramachandran C. Gresser M.J. Tremblay M.L. Kennedy B.P. Science. 1999; 283: 1544-1548Crossref PubMed Scopus (1926) Google Scholar, 4Klaman L.D. Boss O. Peroni O.D. Kim J.K. Martino J.L. Zabolotny J.M. Moghal N. Lubkin M. Kim Y.-B. Sharpe A.H. Stricker-Krongrad A. Shulman G.I. Neel B.G. Kahn B.B. Mol. Cell. Biol. 2000; 20: 5479-5489Crossref PubMed Scopus (1130) Google Scholar). Similar results were also observed when an antisense oligonucleotide was injected into mice (5Zinker B.A. Rondinone C.M. Trevillyan J.M. Gum R.J. Clampit J.E. Waring J.F. Xie N. Wilcox D. Jacobson P. Frost L. Kroeger P.E. Reilly R.M. Koterski S. Opgenorth T.J. Ulrich R.G. Crosby S. Butler M. Murray S.F. McKay R.A. Bhanot S. Monia B.P. Jirousek M.R. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 11357-11362Crossref PubMed Scopus (404) Google Scholar). These compelling biological results coupled with the wealth of structural data, which have been generated since the crystal structure of PTP1B was determined in 1994 (6Barford D. Flint A.J. Tonks N.K. Science. 1994; 263: 1397-1404Crossref PubMed Scopus (687) Google Scholar), have contributed to the rapid design of many potent inhibitors (7Johnson T.O. Ermolieff J. Jirousek M.R. Nat. Rev. Drug Discovery. 2002; 1: 696-709Crossref PubMed Scopus (567) Google Scholar, 8Blaskovich M.A. Kim H.-O. Expert Opin. Ther. Pat. 2002; 12: 871-905Crossref Scopus (66) Google Scholar, 9Liu G. Curr. Med. Chem. 2003; 10: 1407-1421Crossref PubMed Scopus (67) Google Scholar, 10Taylor S.D. Hill B. Exp. Opin. Investig. Drugs. 2004; 13: 199-214Crossref PubMed Scopus (88) Google Scholar, 11Pei Z. Liu G. Lubben T.H. Szczepankiewicz B.G. Curr. Pharm. Des. 2004; 10: 3481-3504Crossref PubMed Scopus (63) Google Scholar, 12Costantino L. Barlocco D. Curr. Med. Chem. 2004; 11: 2725-2747Crossref PubMed Scopus (14) Google Scholar, 13Bialy L. Waldmann H. Angew. Chem. Int. Ed. 2005; 44: 3814-3839Crossref PubMed Scopus (388) Google Scholar). Unfortunately, poor physicochemical properties have limited their development as drug candidates. Crystal structures of PTP1B in complex with phosphorylated peptides corresponding to the IR kinase activation segment (14Salmeen A. Andersen J.N. Myers M.P. Tonks N.K. Barford D. Mol. Cell. 2000; 6: 1401-1412Abstract Full Text Full Text PDF PubMed Scopus (401) Google Scholar) and an autophosphorylated site of the epidermal growth factor receptor (15Jia Z. Barford D. Flint A.J. Tonks N.K. Science. 1995; 268: 1754-1758Crossref PubMed Scopus (558) Google Scholar) reveal that the highly negatively charged substrates bind to multiple positively charged sites (Fig. 1a). The catalytic site, which is located at the base of a shallow pocket called the primary phosphate-binding pocket or A site, is the most polar region of the protein, and it contains the phosphate-binding signature motif (Cys215-Arg221) common to all members of the protein-tyrosine phosphatase family (16Andersen J.N. Mortensen O.H. Peters G.H. Drake P.G. Iversen L.F. Olsen O.H. Jansen P.G. Andersen H.S. Tonks N.K. Møller N.P.H. Mol. Cell. Biol. 2001; 21: 7117-7136Crossref PubMed Scopus (603) Google Scholar). A second phosphate-binding pocket (B site), adjacent to the A site, was identified from the crystal structures of the protein in complex with a small aryl phosphonate (17Puius Y.A. Zhao Y. Sullivan M. Lawrence D.S. Almo S. Zhang Z.-Y. Prot. Natl. Acad. Sci. U. S. A. 1997; 94: 13420-13425Crossref PubMed Scopus (404) Google Scholar) and a bisphosphorylated peptide (14Salmeen A. Andersen J.N. Myers M.P. Tonks N.K. Barford D. Mol. Cell. 2000; 6: 1401-1412Abstract Full Text Full Text PDF PubMed Scopus (401) Google Scholar). The B site, which is larger and shallower than the A site and has lower binding affinity for aryl phosphates, is noncatalytic but may play an important role in determining substrate specificity. A third phosphate-binding site (C site) was identified when the distal Tyr(P) mimetic of a bisphosphonate-containing inhibitor unexpectedly bound in a large flat region of the protein near Lys41 and Arg47 (18Jia Z. Ye Q. Dinaut A.N. Wang Q. Waddleton D. Payette P. Ramachandran C. Kennedy B. Hum G. Taylor S.D. J. Med. Chem. 2001; 44: 4584-4594Crossref PubMed Scopus (72) Google Scholar). Overall, the active site of PTP1B possesses very few desirable drug-design features. The highly charged A site and flat, solvent-exposed B and C sites significantly increase the difficulty of designing potent inhibitors with acceptable membrane permeability. In the past 10 years, many anions that mimic the interactions of phosphate, in the A site, have been incorporated into the design process (19Burke Jr., T.R. Lee K. Acc. Chem. Res. 2003; 36: 426-433Crossref PubMed Scopus (119) Google Scholar). The most potent include the following: oxalylaminothiophene carboxylic acid (OTCA), carboxymethoxybenzoic acid (CMBA), oxalylaminobenzoic acid, and difluoromethylphosphonic acid (DFMP) (Fig. 1b). Unfortunately, the resulting compounds have poor membrane permeability because their highly charged phosphonate mimetics are ionized at physiological pH. We have recently reported the design of a novel five-membered isothiazolidinone (IZD) heterocyclic phosphonate mimetic that possesses only a single delocalized negative charge (Fig. 1b) (20Combs A.P. Yue E.W. Bower M. Ala P.J. Wayland B. Douty B. Takvorian A. Polam P. Wasserman Z. Zhu W. Crawley M. Pruitt J. Sparks R. Glass B. Modi D. McLaughlin E. Bostrom L. Li M. Galya L. Blom K. Hillman M. Gonneville L. Reid B. Wei M. Becker-Pasha M. Klabe R. Huber R. Li Y. Hollis G. Burn T.C. Wynn R. Liu P. Metcalf B. J. Med. Chem. 2005; 48: 6544-6548Crossref PubMed Scopus (96) Google Scholar). IZD-containing compounds are potent competitive inhibitors of PTP1B, but more importantly, they have been shown recently to have cellular activity in an IR phosphorylation assay (21Combs A.P. Zhu W. Crawley M.L. Glass B. Polam P. Sparks R.B. Modi D. Takvorian A. McLaughlin E. Yue E.W. Wasserman Z. Bower M. Wei M. Rupar M. Ala P.J. Reid B.M. Ellis D. Gonneville L. Emm T. Taylor N. Yeleswaram S. Li Y. Wynn R. Burn T.C. Hollis G. Liu P.C.C. Metcalf B. J. Med. Chem. 2006; 49: 3774-37789Crossref PubMed Scopus (97) Google Scholar). The IZD ring was specifically designed to incorporate all previously observed binding features of known phosphonate mimetics. For example, it simultaneously mimics the interactions of the three phosphate oxygens of DFMP, which is arguably the most potent phosphonate mimetic, and the carboxymethyl group of CMBA (Fig. 2a). Combining these features into a single heterocycle has produced one of the most potent phosphonate mimetics reported to date. One of our most potent IZD-containing peptides (compound 8) has an IC50 value of 190 nm, which is ∼9-fold more potent than the equivalent DFMP-containing analog (Table 3). In order to gain insights into the structural basis for inhibition of PTP1B by IZD-containing compounds, we have solved the crystal structures of the protein in complex with IZD-phenol and several IZD-containing peptides (Table 2). Identifying the key binding interactions should aid the design of small molecule inhibitors as well as the design of even more potent, less polar phosphate mimetics.TABLE 3Phosphonate mimetic SAR Open table in a new tab TABLE 2Data collection and refinement statistics for PTP1B in complex with phosphonate mimetics Open table in a new tab Two constructs of the catalytic domain of human PTP1B (435 amino acids), 1-321 and PTP1B-(1-His6-298), were expressed in Escherichia coli essentially as described previously (22Barford D. Keller J.C. Flint A.J. Tonks N.K. J. Mol. Biol. 1994; 239: 726-730Crossref PubMed Scopus (48) Google Scholar). Briefly, BL21(DE3)-competent cells (Novagen) were transformed with the expression plasmids as described by the manufacturer. LB broth (2 ml) containing 100 μg/ml ampicillin was inoculated with a single colony from an ampicillin-containing LB agar plate and incubated in a 15-ml tube for 2 h of shaking at 225 rpm at 37 °C. The culture was then used to inoculate 50 ml of Terrific Broth (Media Tech) containing 100 μg/ml ampicillin in a 500-ml vented Erlenmeyer flask. The following day, the 50-ml culture was used to inoculate 1 liter of Terrific Broth/ampicillin in a 3-liter vented Fernbach culture flask. When the A600 reached 1, 0.5 mm isopropyl β-d-thiogalactopyranoside was added, and the culture was incubated for 16 h of shaking at 75 rpm at 27 °C. The culture was pelleted at 5,800 × g, immediately frozen on dry ice, and stored at -80 °C. Cell paste was suspended (5 ml/g of cells) at 4 °C in buffer A (20 mm Tris, pH 7.5, 150 mm NaCl, 10% glycerol, 100 μm leupeptin, 1 μm pepstatin A, and 1 mm phenylmethylsulfonyl fluoride (Sigma)) and stirred for 15 min. The mixture was then homogenized for 30 s using an Ultra-Turrax T 25 basic (IKA Works, Inc.), lysed using an EmulsiFlex-C5 homogenizer (Avestin, Inc.) with two passes at 10,000-15,000 p.s.i., and then centrifuged at 11,000 × g for 30 min. The supernatant containing soluble PTP1B-(1-His6-298) was mixed with Talon metal affinity resin (BD Biosciences) at a protein to resin ratio of 10:1 (mg/ml) and stirred for 2 h. The protein/resin mixture was then poured into a column and washed with buffer A until the eluate A280 < 0.1. The resin was then washed with 2 column volumes of buffer A containing 10 mm imidazole, and the protein was eluted with buffer A containing 100 mm imidazole. The supernatant containing soluble PTP1B-(1-321) was mixed with Q-Sepharose Fast Flow resin (GE Healthcare) and placed in a spinner flask overnight at 4 °C. The protein/resin mixture was then filtered through a sintered glass filter, resuspended in buffer B (20 mm Tris, pH 7.5, and 1 mm EDTA), poured into a column, and washed until A280 < 0.1. The protein was eluted with a linear gradient of NaCl (0-500 mm) in buffer B using a total of 10 column volumes. Column fractions were analyzed by SDS-PAGE, and those containing PTP1B were pooled and mixed with diammonium sulfate (4 m stock) to a final concentration of 1.7 m. The mixture was then diluted with an equal volume of buffer B containing 1.7 m diammonium sulfate and loaded onto a fast flow phenyl-Sepharose column (GE Healthcare). The protein was eluted with a decreasing diammonium sulfate gradient (1.7-0 m) using 8 column volumes. Column fractions from both preparations were analyzed by SDS-PAGE, and those containing PTP1B were concentrated to 5 mg/ml using a stirred cell with a YM-10 membrane (Millipore). The concentrated protein (2-ml aliquots) was injected onto a HiLoad Superdex 75 26/60 Prep Grade column connected to an ÁKTApurifier chromatography system (Amersham Biosciences). The column was pre-equilibrated with buffer C (10 mm Tris, pH 7.5, 25 mm NaCl, and 1 mm EDTA) and run at a flow rate of 2.5 ml/min. Fractions containing pure protein were concentrated to 10 mg/ml and stored in 200-μl aliquots at -20 °C. Approximately 20 mg of pure PTP1B was obtained from 12 g of cells. PTP1B enzymatic assays were performed essentially as described previously (23Szczepankiewicz B.G. Liu G. Hajduk P.J. Abad-Zapatero C. Pei Z. Xin Z. Lubben T.H. Trevillyan J.M. Stashko M.A. Ballaron S.J. Liang H. Huang F. Hutchins C.W. Fesik S.W. Jirousek M.R. J. Am. Chem. Soc. 2003; 125: 4087-4096Crossref PubMed Scopus (225) Google Scholar). Briefly, protein activity was determined by measuring the rate of hydrolysis of p-nitrophenyl phosphate. The reaction was performed at room temperature in 25 mm Tris-bis-propane, pH 7.0, containing 5 nm PTP1B, 0.1 mg/ml bovine serum albumin, and 1 mm p-nitrophenyl phosphate, which is its Km value. IC50 values were determined by fitting the initial rates of p-nitrophenol production to a sigmoidal dose-response equation using Prism 3.0 (GraphPad Software). The two protein constructs, PTP1B-(1-321) and PTP1B-(1-His6-298), were crystallized in the presence and absence of inhibitors (dissolved in water pH 7.0). All crystals were grown using the vapor diffusion method with or without microseeding in 24- or 96-well plates in which the protein drops were equilibrated against 1 or 0.15 ml of well solution, respectively, at 4 or 10 °C. Crystals of PTP1B-(1-321)—The apoprotein crystallized in the trigonal space group P3121 as hexagonal rods (0.3 × 0.2 × 0.01 mm) 7 days after mixing 2 μl of protein (10 mg/ml) with 2 μl of well solution (100 mm Hepes, pH 6.6, 14-16% PEG 8000, and 200 mm magnesium acetate) at 4 °C. Crystals of PTP1B in complex with compounds 1, 2, and 3 were grown by mixing 2 μl of protein/inhibitor solution (10 mg/ml protein with 3, 10, and 2.4 mm inhibitor, respectively) with 2 μl of well solution (100 mm Hepes, pH 6.6, 14-16% PEG 8000, and 200 mm magnesium acetate). Hexagonal rods containing compounds 1 (0.06 × 0.06 × 0.12 mm), 2 (0.02 × 0.02 × 0.25 mm), and 3 (0.02 × 0.02 × 0.25 mm) appeared 7-10 days after macroseeding at 4 °C. Crystals of PTP1B 1-[His6]-298—The apoprotein crystallized in the orthorhombic P212121 and monoclinic C2 space groups. The orthorhombic crystals grew into thin plates (0.3 × 0.2 × 0.01 mm) 7 days after mixing 1 μl of protein solution (10 mg/ml protein, 0.025% β-octyl glucoside, and 2 mm TCEP) with 1 μl of well solution (100 mm sodium/potassium phosphate, pH 5.9, and 25-35% methylpentanediol) at 4 °C. The monoclinic crystals grew into large bipyramids (0.2 mm in length) 3-5 days after mixing 2 μl of protein (10 mg/ml) with 2 μl of well solution (12-16% PEG 3000, 100 mm Hepes, pH 7.0-8.0, 200 mm magnesium acetate, and 2 mm TCEP) and microseeding at 4 °C. Rod-shaped crystals of PTP1B/4 (0.04 × 0.04 × 0.25 mm) appeared 7-10 days after mixing 4 μl of protein/inhibitor solution (10 mg/ml protein, 0.025% β-octyl glucoside, 2 mm TCEP, and 2 mm compound) with 4 μl of well solution (100 mm Hepes, pH 8.5, and 1.12 m sodium citrate) and microseeding at 10 °C. Rod-shaped crystals of PTP1B/5 (0.05 × 0.05 × 0.3 mm) appeared 10 days after mixing 2 μl of protein/inhibitor solution (10 mg/ml protein, 0.025% β-octyl glucoside, 2 mm TCEP, and 2.4 mm compound) with 2 μl of well solution (100 mm Tris, pH 8.5, and 1.4-1.8 m ammonium sulfate) at 4 °C. Crystals were transferred stepwise at 1-min intervals into cryoprotectant solutions of mother liquor containing 10 and 20% glycerol and then flash-cooled in a nitrogen stream at -180 °C. For PTP1B/1 crystals, additional compound (10 mm) was added to the cryoprotectant solution to prevent the inhibitor from soaking out during the procedure. The orthorhombic apocrystals were cooled in mother liquor without additional cryoprotectant. Diffraction data were collected with one crystal per data set on a Rigaku/MSC RAXIS IV++ imaging system mounted on a Rigaku/MSC MicroMax™-007 rotating anode microfocus generator (CuKα) operating at 40 kV and 20 mA and equipped with a 0.3-mm cathode, 0.5- and 0.03-mm double collimator, and a Blue Confocal Max-Flux® optical system. Diffraction intensities were integrated and scaled using Crystal-Clear (Rigaku/MSC). The orthorhombic and monoclinic apostructures and the structures containing 2 and 4 were solved by molecular replacement using the protein coordinates from the Protein Data Bank (PDB) (24Berman H.M. Westbrook J. Feng Z. Gilliland G. Bhat T.N. Weissig H. Shindyalov I.N. Bourne P.E. Nucleic Acids Res. 2000; 28: 235-242Crossref PubMed Scopus (27906) Google Scholar) entry code 1EE0 as the search model and the program CNX (Accelrys). All rotation and translation searches yielded single consistent solutions for data between 15 and 4 Å. The resulting models were subjected to rigid body refinement followed by simulated annealing (using torsion angle dynamics) and positional and temperature factor refinements using CNX. All other structures were solved by difference Fourier methods using the relevant starting model. The conformations of the bound inhibitors were unambiguously determined after one or two cycles of refinement. The final models were generated after several cycles of model building and refinement using QUANTA (Accelrys) and CNX, respectively. Only the (S)-IZD stereoisomers of 2 and 3 were observed in the crystal structures even though diastereoisomeric mixtures were used in crystallization experiments. Solvent molecules were added by visual inspection of electron density maps using X-SOLVATE (Accelrys). The final models are consistent with simulated annealed omit maps: 1 (Fig. 2b), apo P212121 (Fig. 3a), 4 (Fig. 7a), and 2 (Fig. 8b). Crystallographic data and refinement statistics are listed in Tables 1 and 2.FIGURE 7Crystal structure of PTP1B/4. a, stereo view of the 2Fo - Fc simulated annealed omit map (contoured at 1σ) and atomic model for PTP1B/4. The inhibitor was omitted from the model prior to a cycle of simulated annealing and was not used in the calculation of phases. Note the distal IZD-phenyl moiety is the least well defined region of the inhibitor. b, superposition of PTP1B/4 (surface and stick bonds) and 1 (ball-and-stick) reveals that the peptide scaffold of 4 does not alter the bound conformation of the IZD-phenyl moiety. Dashed lines indicate hydrogen bonds between 4 and backbone nitrogen of Arg47 and side chain of Asp48.View Large Image Figure ViewerDownload Hi-res image Download (PPT)FIGURE 8a, superposition of PTP1B/3 (ball-and-stick and thick bonds) and 4 (thin bonds) reveals that the bound conformation of the peptide scaffold is very similar whether or not the heterocycle is saturated or unsaturated. The only differences are due to crystal contacts (not shown) in the PTP1B/4 (compare benzyl substituents) and structural differences between the inhibitors. b, stereo view of the 2Fo - Fc simulated annealed omit map (contoured at 1σ) and atomic model for PTP1B/2. The inhibitor was omitted from the model prior to a cycle of simulated annealing and was not used in the calculation of phases. The entire inhibitor is very well defined except for the benzyl group, which binds between Arg47 and Asp48. Note only the (S)-IZD isomer of 2 crystallized in complex with PTP1B, even though a diastereoisomeric mixture was present during crystallization. c, superposition of PTP1B/2 (surface and thick and thin bonds) and 5 (ball-and-stick) reveals that IZD and DFMP project the peptide scaffold in exactly the same position in the C site. Note Wat4 hydrogen bonds to DFMP, and a sulfate ion in the B site hydrogen bonds to Arg24 and Arg254 in PTP1B/5; the protein of the latter was omitted for clarity.View Large Image Figure ViewerDownload Hi-res image Download (PPT)TABLE 1Data collection and refinement statistics for apoPTP1BProtein constructPTP1B 1-(His6)-298PTP1B 1-321Data collection statisticsTemperature (°C)−1804−180−180Space groupP212121P212121C2P3121Unit cell (Å, °)a = 40.6a = 41.1a = 106.8a = b = 88.3b = 69.2b = 70.4b = 95.5c = 104.1c = 100.8c = 102.4c = 70.7β = 109.7Resolution range (Å)aValues in parentheses are for the highest resolution bin.33-1.5 (1.55-1.5)19-2.0 (2.07-2.0)27-2.1 (2.18-2.1)44-2.1 (2.18-2.1)Molecules/absorbance unit1121No. of unique reflections44,32718,00338,99127,186Average redundancy4.0 (2.3)2.5 (2.4)3.7 (3.7)5.0 (5.0)Completeness (%)95.7 (73.0)86.9 (89.6)99.8 (100)97.4 (95.3)〈I/σ 〉bAverage of the diffraction intensities divided by their standard deviations.16.5 (3.0)5.1 (1.8)14.8 (5.6)17.6 (4.6)Rmerge (%)cRmerge = ∑hkl∑i|Ihkl,i - 〈Ihkl 〉|/∑hkl∑i(Ihkl,i), where 〈Ihkl 〉 is the average intensity of the multiple Ihkl,i observations.4.9 (32.4)10.3 (41.1)5.8 (13.7)5.2 (31.0)Refinement statisticsNo. of protein atoms2341227244742410No. of solvent atoms341127507207No. of heteroatoms16027Rmsd, bond (Å/°)0.011/1.540.008/1.400.008/1.40.011/1.54Resolution range (Å)33-1.519-2.027-2.144-2.1Rfactor/Rfree (%)dRfactor = ∑hkl||Fo| - |Fc||/∑hkl|Fo|, where Fo and Fc are observed and calculated structure factor amplitudes for reflection hkl, respectively. Rfree is the same as Rfactor for a random 10% of reflections excluded from refinement.20.1/23.721.8/27.723.9/29.822.8/27.3Resolution in final structure[2xHis]-2823-2807-2803-297PDB ID2CM2NAeNA indicates not applicable.2CM3NAa Values in parentheses are for the highest resolution bin.b Average of the diffraction intensities divided by their standard deviations.c Rmerge = ∑hkl∑i|Ihkl,i - 〈Ihkl 〉|/∑hkl∑i(Ihkl,i), where 〈Ihkl 〉 is the average intensity of the multiple Ihkl,i observations.d Rfactor = ∑hkl||Fo| - |Fc||/∑hkl|Fo|, where Fo and Fc are observed and calculated structure factor amplitudes for reflection hkl, respectively. Rfree is the same as Rfactor for a random 10% of reflections excluded from refinement.e NA indicates not applicable. Open table in a new tab To identify the unique structural features responsible for the high potency of IZD-containing compounds, we have determined the crystal structure of PTP1B in complex with the small unsaturated IZD-phenol 1, which is a potent inhibitor for its size (IC50 = 1.35 μm) (Table 2). Crystal structures of PTP1B in complex with unsaturated and saturated IZD- and DFMP-containing peptides were also determined (Table 2) to explain why the saturated IZD 8 is 9- and 16-fold more potent than the equivalent DFMP 5 and unsaturated 6 analogs (Table 3), respectively, as well as identify key interactions outside of the A site that might aid the design of non-peptidic inhibitors. IZD Is Highly Complementary to the Primary Phosphate Binding Pocket—The structure of PTP1B/1 reveals extensive interactions between the inhibitor and the A site. The heterocycle binds at the center of the phosphate-binding loop (Cys215-Arg221), where its four hydrogen bonding acceptors are in the vicinity (2.7-3.3 Å) of 10 donors (Fig. 2c). The sulfone oxygens hydrogen bond to N-η of Arg221 and the backbone nitrogens of Ser216, Ala217, Gly218, Ile219, and Gly220. The nitrogen anion interacts with the helix dipole of α4 (Arg221-Arg238) and hydrogen bonds to N-ϵ and backbone NH of Arg221, and the carbonyl oxygen hydrogen bonds to the backbone nitrogen of Phe182 and side chain of Gln266. An additional hydrogen bond is observed between the side chain of Asp181 and the π electrons of the unsaturated IZD. The hydrophobic side of the heterocycle and phenyl ring directly attached to it interact with the hydrophobic half of the A site as follows: Tyr46, Val49, Ala217, Ile219, Phe182, and the aliphatic portion of the side chain of Gln262. The latter van der Waals interactions and the extensive network of hydrogen bonds are key reasons why IZD-phenol is a potent inhibitor. IZD Displaces All Tightly Bound Waters in the A Site—The recently reported high resolution crystal structure of the apoprotein in the P3121 space group revealed the presence of three water molecules that bind in the same locations as the phosphonate oxygens of Tyr(P)-containing substrates (25Pedersen A.K. Peters G.H. Møller K.B. Iversen L.F. Kastrup J.S. Acta Crystallogr. Sect. D Biol. Crystallogr. 2004; 60: 1527-1534Crossref PubMed Scopus (49) Google Scholar). To confirm that the observed water structure is independent of crystallization conditions and space group, we solved the structures of the apoprotein in two additional space groups, P212121 (Fig. 3a) and C2, as well as the reported P3121. Four tightly bound water molecules (Wat1-4) are observed in each structure, except the C2 structure that does not contain Wat4 (Fig. 3b). Three of the waters hydrogen bond to backbone nitrogens of the phosphate-binding loop: Wat1 hydrogen bonds to Arg221, Wat2 to Ile219 and Gly220, and Wat3 to Ser216 and Ala217. The fourth water hydrogen bonds to the side chain of Gln266. Other waters in the A site are not conserved between the structures and thus probably interact weakly with the protein. The second aryl phosphate-binding site (B site) is also highly solvated: five conserved waters are located near Arg254. The most tightly bound water, based on electron density, is Wat5, which hydrogen bonds directly to the side chain Arg254 and appears to organize the surrounding water. Importantly, the presence of Wat2, Wat3, and most of the conserved B site waters does not appear to be an artifact of the low temperature data collection, because they are also present in the P212121 apostructure solved at 4 °C. Finally, our trigonal structure is identical to the reported structure, containing a glycerol in the hydrophobic half of the A site and a mixture of opened and closed conformations of the flap (Phe180-Ile184). The disordered conformation of the flap and the presence of glycerol, however, do not appear to influence the water structure associated with the phosphate-binding loop because the waters are also present in our P212121 and C2 apostructures that only contain the opened conformation of the flap. A superposition of apoPTP1B (P212121) and PTP1B in complex with the unsaturated IZD 4 (Table 2; our highest resolution structure) reveals that the IZD-phenyl moiety displaces at least nine waters, more then any other Tyr(P) mimetic (Fig. 4). The sulfone oxygens and nitrogen anion displace Wat1-Wat3; the carbonyl oxygen and phenyl each displace two waters, and the side chain of Arg221 which shifts as a result of ligand binding displaces two waters. The three phosphate oxygens of the DFMP moiety of 5 displace Wat1-Wat3 but not Wat4, which shifts by 1.5 Å to accommodate one of the fluorine atoms and remains trapped under the flap where it hydrogen bonds to the fluorine atoms of DFMP, the backbone NH of Phe182, and side chain of Gln266. The CMBA moiety of the PDB entry code 1G7G only displaces Wat1 and Wat4, and the OTCA of 1C83 displaces Wat1-Wat3 but traps Wat4. So far IZD and the recently reported thiadiazolidinone (26Black E. Breed J. Breeze A.L. Embrey K. Garcia R. Gero T.W. Godfrey L. Kenny P.W. Morley A.D. Minshull C.A. Pannifer A.D. Read J. Rees A. Russell D.J. Toader D. Tucker J. Bioorg. Med. Chem. Lett. 2005; 15: 2503-2507Crossref PubMed Scopus (72) Google Scholar) heterocycles are the only phosphonate mimetics that possess a single negative charge and displace all waters in the A site of the closed conformation of the protein. IZD-Phenol Induces the Closed Conformation of the Protein—To identify conf
Referência(s)