PEDS Advance Access published online on May 2, 2008
Protein Engineering Design and Selection, doi:10.1093/protein/gzn017
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Construction and characterization of a fully active PXR/SRC-1 tethered protein with increased stability
1Structural Chemistry Department 2Drug Metabolism Department, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
3 To whom correspondence should be addressed. E-mail: wenyan.wang{at}spcorp.com, charles.lesburg{at}spcorp.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
The nuclear xenobiotic receptor PXR is a ligand-inducible transcription factor regulating drug-metabolizing enzymes and transporters and a master switch mediating potentially adverse drug–drug interactions. In addition to binding a coactivator protein such as SRC-1, the C-terminal ligand-binding domain (LBD) is solely responsible for ligand recognition and thus the ligand-dependent downstream effects. In an effort to facilitate structural studies of PXR to understand and abolish the interactions between PXR and its ligands, several recombinant PXR/SRC-1 constructs were designed and evaluated for expression, stability and activity. Expression strategies employing either dual expression or translationally coupled bicistronic expression were found to be unsuitable for producing stable PXR in a stochiometric complex with a peptide derived from SRC-1 (SRC-1p). A single polypeptide chain encompassing PXR and SRC-1p tethered with a peptidyl linker was designed to promote intramolecular complex formation. This tethered protein was overexpressed as a soluble protein and required no additional SRC-1p for further stabilization. X-ray crystal structures in the presence and absence of the known PXR agonist SR-12813 were determined to high resolution. In addition, a circular dichroism-based binding assay was developed to allow rapid evaluation of PXR ligand affinity, making this tethered protein a convenient and effective reagent for the rational attenuation of drug-induced PXR-mediated metabolism.
Keywords: circular dichroism/protein engineering/PXR/SRC-1/stability/tethered protein
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Nuclear receptors (NRs) are a eukaryotic superfamily of ligand-inducible transcription factors regulating development and metabolism. Pregnane X receptor (PXR, NR1I2) is a NR evolved to protect the body from the toxicity of xenobiotics. PXR can be activated by a variety of structurally unrelated chemicals which induce expression of multiple phase I and phase II drug-metabolizing enzymes and transporters. It has been shown that this pathway, which includes cytochrome P450 (CYP) enzymes, is responsible for metabolism and elimination of more than half of all clinical drugs, therefore giving rise to many drug safety and drug interaction issues related to pharmaceutical development (Moore and Kliewer, 2000
; Willson et al., 2002; Honkakoski, 2003).
PXR was originally identified based on upregulation of CYP3A isozymes upon exposure to a structurally diverse repertoire of chemicals, including several widely used prescription drugs such as dexamethasone, RU486, corticosterone, estradiol and dihydrotestosterone (Blumberg et al., 1998; Kliewer et al., 1998). PXR contains two modular domains common to the NR superfamily—a highly conserved DNA-binding domain (DBD) located at the N-terminus and a more divergent ligand-binding domain (LBD) at the C-terminus. Containing the binding determinants for both ligands and coactivators or corepressors, the LBD of PXR is critical to understanding its function (Weatherman et al., 1999; Wang and LeCluyse, 2003) and perhaps the basis of ligand-mediated transcriptional effects. Like many other transcription factors, PXR is a protein of inherently low stability. The nascent PXR is located in the cytosol, complexed with and silenced by a corepressor. Upon ligand binding, the repressor is displaced by a coactivator such as steroid receptor coactivator 1 (SRC-1); the activated PXR forms a heterodimer with another NR RXR and translocates into the nucleus. The paired DBD domains bind to nucleotide hexamer (AGGTTC) repeats located in the promoter regions of the downstream genes, leading to initiation of transcription. This orchestrated mechanism provides precise regulation of the PXR-mediated xenobiotic metabolism (Kliewer et al., 2002; Handschin and Meyer, 2003).
In recent years, significant progress has been made in understanding PXR-mediated CYP induction using cell-based assays (Jones et al., 2002; Raucy et al., 2002) and structural analyses. Recombinant PXR-LBD produced using a dual plasmid expression system has led to several X-ray crystal structures including the apoprotein (Watkins et al., 2001), a binary complex with the PXR agonist hyperforin (Watkins et al., 2003b) as well as a ternary complex with an SRC-1 peptide (SRC-1p) and the PXR agonist SR-12813 (Watkins et al., 2003a). These structures reveal that PXR has a typical NR structure comprised of a three-layer 
-helical sandwich and a five-stranded anti-parallel β-sheet (Moras and Gronemeyer, 1998; Weatherman et al., 1999). The large ligand-binding pocket is highly hydrophobic and expandable, a unique feature contributing to the promiscuity of PXR towards a wide range of ligands of different sizes and chemical properties. The flexibility of the pocket allows ligands to take on several simultaneous binding orientations, restriction of which has been observed by binding a 25-mer coactivator peptide (SRC-1p, 676-CPSSHSSLTERHILHRLLQEGSPS-700) (Watkins et al., 2003a). SRC-1p contains an LxxLL motif that adopts an -helical conformation and binds to the surface of PXR-LBD, forming a ‘charge clamp’ as observed in other NR-coactivator complexes (Darimont et al., 1998; Nolte et al., 1998; Gampe et al., 2000; Xu et al., 2001). This interaction may reduce ‘breathing’ of the cavity and conformationally restrict the ligand to an active orientation, thereby increasing the thermal stability of the protein.
In order to eliminate PXR-ligand interactions, a stable and functional PXR-LBD form is needed to facilitate biophysical characterization and structure determination of PXR-ligand complexes. Our initial effort in reproducing the published results using a recombinant PXR derived from the dual plasmid expression system (Watkins et al., 2001) led to a suboptimal ratio of SRC-1p to PXR, and the proteins appeared to be unstable. In the current report, two novel Escherichia coli expression strategies were explored in an attempt to increase the stability while maintaining the ligand-binding function of PXR. The thermal stability of these recombinant proteins was evaluated using a circular dichroism-based binding assay. X-ray crystal structures of the single-chain hybrid protein, with and without bound ligand, were determined to high resolution.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Unless otherwise noted, all chemicals used in this study were purchased from Sigma (St Louis, MO, USA).
Design of recombinant expression plasmids
The sequences of the oligonucleotide primers used in this report are listed in Table I. The DNA insert in each plasmid has been sequenced to ensure that no mutation occurred during the PCR and cloning procedures.
|
Dual plasmid expression system (D14-PXR and pACYC184-SRC)
PXR-LBD DNA fragment P1 (coding for an N-terminal 6 x His and amino acid residues 130–434) was amplified using the primers H6-hPXR(130)-F and PBC-PXR-R. The PCR product was cloned into vector pENTRTM/SD/D-TOPO® and subsequently recombined into the destination vector pDEST14 to generate D14-PXR following the manufacturer’s protocol (Invitrogen, Carlsbad, CA, USA).
A 320 bp DNA fragment containing T7 promoter region was amplified from vector pRSETa using primers pRSETa-F and pRSETa-R, digested with restriction enzymes BamHI and subsequently ligated between the BamHI and EcoRV sites in vector pACYC184 (ATCC) to construct the vector pACYC184-T7. A PCR fragment encoding human SRC-1623-710 was amplified using primers SRC-Nd and SRC-Ba, digested with BamHI and NdeI and cloned into the BamHI and NdeI sites in pACYC184-T7, resulting in pACYC184-SRC.
Construction of bicistronic expression plasmids
Plasmids containing PXR-LBD and SRC-1p as two cistrons were constructed in pDEST14 vector (Invitrogen). DNA sequences encoding PXR-LBD (PXR130–434) and the SRC-1623–710 in both PXR-SRC and SRC-PXR orientations were designed. A stop codon at the 3' end of the first cistron followed by the T7 gene ribosome binding site (RBS) at the 5' end of the second cistron was inserted to avoid a fusion protein (Wu et al., 1999). The spacer between the two cistrons was minimized in order to achieve tight coupling of translation. A schematic representation of the expression system organization is presented in Fig. 1.
|
D14-PXR-RBS-SRC. PXR-LBD DNA fragment P1 (encoding for an N-terminal 6 x His and amino acid residues 130–434) was amplified using the primers H6-hPXR(130)-F and PBC-PXR-R. SRC-1 DNA fragment S1 (encoding for the amino acid residues 623–710) was amplified using the primers PBC-SRC-F and SRC710-R. Mixture of P1 and S1 at 1:1 molar ratio served as templates for a second round of PCR amplification using primers H6-hPXR(130)-F and SRC710-R. The resulting PCR product was ligated into a Gateway donor vector pENTRTM/SD/D-TOPO® to obtain a plasmid TOPO-PXR-RBS-SRC following the manufacturer’s protocol (Invitrogen, Carlsbad, CA, USA). The expression plasmid D14-PXR-RBS-SRC was subsequently obtained from an LR in vitro recombination reaction by mixing plasmids TOPO-PXR-RBS-SRC and pDEST14 under catalysis of Clonase (Invitrogen, Carlsbad, CA, USA).
D14-SRC-RBS-PXR. SRC-1 DNA fragment S2 (encoding an N-terminal 6 x His and amino acid residues 623–710) was amplified using the primers PSRC-F and PBC-SRC-R; human PXR DNA fragment P2 (encoding the amino acid residues 130–434) was amplified using the primers PBC PXR-F and hPXR(434)-R. Mixture of S2 and P2 served as templates for a subsequent PCR reaction using primers PSRC-F and hPXR(434)-R. The resulting PCR product was cloned into the vector pDEST14 to obtain the expression plasmid D14-SRC-RBS-PXR using the Gateway Cloning Strategy (Invitrogen, Carlsbad, CA, USA).
Construction of tethered expression plasmids (D14-PXR-Ln-SRCp)
A DNA fragment encoding an N-terminal MKKGHHHHHHG sequence, followed by PXR-LBD (residues 130–434), a pentapeptide linker (GGSGG), a peptide fragment of SRC-1 (residues 678-710) and an opal stop codon were amplified using the primers hPXR(130)-F and PL10-SRC-R. The resulting PCR fragment was cloned into vector pDEST14 to obtain the expression plasmid D14-PXR-L10-SRC1p (the number following L indicates the linker length in amino acid residues) using the Gateway cloning strategy as described earlier (Invitrogen, Carlsbad, CA, USA). Plasmids D14-PXR-L8-SRC1p and D14-PXR-L13-SRC1p were obtained by site-directed mutagenesis using primer pairs L8-F/L8-R and L13-F/L13-R, respectively.
A single colony of freshly transformed E.coli BL21(DE3) cells was used to initiate growth in Terrific Broth medium supplemented with 0.2% glucose and 100 µg/ml ampicillin (plus 34 µg/ml chloramphenicol for the dual plasmid system) at 37°C until the cell density reached an OD600 of 3–4, the fermentation temperature was rapidly lowered to 16°C and recombinant protein expression was induced upon addition of 0.2 mM isopropyl β-D-thiogalactopyranoside. The cells were harvested 16 h after induction and frozen at –20°C prior to purification.
Cell pellets from 2 l of fermentation were resuspended in 100 ml lysis buffer containing 25 mM HEPES, pH 7.9, 5% v/v glycerol, 150 mM NaCl, 1 mM DTT, 10 mM imidazole, EDTA-free Protease Inhibitors (Roche Diagnostic, Indianapolis, IN, USA) and 5000 units/l benzonase (Sigma, St Louis, MO, USA) using a Dounce tissue homogenizer (Bellco Glass, NJ, USA) and subsequently disrupted by two passes through a microfluidizer (Model M-110F, Microfluidics, MA, USA) operated at 10 000 psi. The lysate was clarified by centrifugation at 20 000x g for 60 min. The supernatant was applied at 20 ml/min to 5 ml of Ni-NTA resin (Qiagen) in the presence of 10 mM imidazole. The column was washed with 15–20 column volumes (CVs) of lysis buffer. The bound PXR protein was eluted with 5 CVs of lysis buffer supplemented with 250 mM imidazole.
The eluted fractions containing PXR was applied to 10 ml of S-Sepharose Fast Flow resin (Amersham Pharmacia) equilibrated with buffer A (25 mM HEPES, pH 7.9, 5% v/v glycerol, 150 mM NaCl, 5 mM DTT). The flow-through was collected, concentrated and applied to a Superdex 200 column (20 x 60 cm, Amersham Pharmacia) pre-equilibrated with buffer A. Fractions containing greater than 95% pure recombinant PXR protein, as judged by SDS–PAGE and N-terminal peptide sequencing, were pooled and flash-frozen in liquid nitrogen prior to storage at –80°C.
Thermal denaturation monitored by circular dichroism (TdCD)
Circular dichroism (CD) spectra were recorded in the wavelength range 200–300 nm in buffer A at 296 K on a Jasco J-810 spectrapolarimeter. Spectra were acquired at a protein concentration of 2.5 µM using a 1 mm cuvette with a slit width set to 2 nm and a response time of 1 s. In the drug-binding studies, 25 µM compound was incubated with the protein for 10 min prior to data acquisition.
CD-monitored thermal denaturation was carried out in buffer A. Thermal scans were performed in a 1 mm cuvette, following the ellipticity at 220 nm using a response time of 4 s. A PTC-424S six position automatic Peltier accessory allowed continuous monitoring of the thermal transition at a constant rate of 2°C/min. The data were analyzed using the JASCO software assuming a two-state reversible equilibrium transition (Buczek et al., 2002).
Induction of human hepatocytes
Cryopreserved human hepatocytes (lot: ETA and ONQ) were purchased from In Vitro Technologies (Baltimore, MD, USA). Hepatocytes were thawed in a 37°C water bath and transferred immediately to warm InVitroGRO CP medium. The cell suspension was diluted to 0.7 x 106 viable cells/ml with InVitroGRO CP medium. Cells were seeded in type I collagen-coated 48-well tissue culture plates. To each well, 0.2 ml of the cell suspension was added. The plate was placed into a 37°C, 5% CO2 incubator overnight to allow the cells to attach. Afterwards, cells were treated for 3 days with DMSO (vehicle control), and with the following chemicals at concentrations of 3, 10, 30 and 50 µM: clotrimazole, pregnenolone 16--carbonitrile (PCN), troglitazone, troleandomycin, sulfinpyrazole, and with hyperforin at 0.07, 0.21, 0.62 and 1.86 µM in an InVitroGRO HI medium. At day 5, cells were harvested and mRNA levels of CYP 3A4 was determined using real-time PCR (Cui et al., 2005).
Crystallization and diffraction data collection
Crystals of PXR-LBD-linker-SRC1p (with linkers of various lengths reported here) were grown using the hanging-drop vapor diffusion method in which 1 µl of protein (10 mg/ml) in buffer A was mixed with an equal volume of precipitant and sealed in close proximity to 1 ml of the precipitant solution and allowed to equilibrate at 4°C. The precipitant solution contained between 10 and 30% (v/v) 2,4-methylpentanediol (MPD) as well as 100 mM imidazole/HCl buffer pH 8; isopropanol could be substituted for MPD as the precipitant. Crystals were flash-cooled in liquid nitrogen after soaking in an artificial mother liquor containing 30% (v/v) MPD and 100 mM imidazole/HCl pH 8.0. For the complex with SR-12813, crystals were grown using protein incubated with 1 mM compound in 1% (v/v) DMSO for 1 h prior to dispensing the crystallization drops. Diffraction data were collected using a Rigaku R-AXIS HTC detector mounted on a Rigaku FR-E SuperBright X-ray generator equipped with VariMax HF X-ray focusing optics. Data were collected, indexed, integrated, scaled, and reduced using the program CrystalClear 1.3.6 (Rigaku, The Woodlands, TX, USA). The structure was solved using molecular replacement as implemented in the program MOLREP/CCP4 (Collaborative Computational Project No. 4, 1994) using a monomer of PXR-LBD from PDB code 1NRL
[PDB]
as a molecular replacement probe. Coordinate refinement and electron density map calculations were performed using autoBUSTER (Tronrud, 1987; Bricogne, 1993; Bricogne and Irwin, 1996; Bricogne, 1997; Roversi et al., 2000) and CCP4 (Collaborative Computational Project No. 4, 1994), respectively. Model and map visualizations were performed using O (Jones et al., 1991); cavity calculations were performed using VOIDOO (Kleywegt and Jones, 1994). Figures were prepared using PyMOL (DeLano, 2002). Coordinates and structure factors have been deposited with PDB codes 3ctb and 3ctc.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Expression of PXR in the presence and absence of the SRC-1 fragment
Expression of PXR-LBD in E. coli has resulted in rapid protein degradation, and it was reported that coexpression of SRC-1p greatly enhanced the solubility and stability of the protein (Jones et al., 2002). To confirm this observation, D14-PXR was transformed into E. coli BL21(DE3) alone, or together with the plasmid pACYC184-SRC. In the absence of exogenous SRC-1p, PXR was not detected in either the cell lysate or Ni-NTA eluate (Fig. 2, Lane 1 and 2). In comparison, when co-expressed with SRC-1p, protein bands corresponding to both the PXR (
36 kDa) and SRC-1 (9 kDa) were identified (Fig. 2, Lane 7 and 8). As only the recombinant PXR contained a polyhistidine tag, the co-purified SRC-1p must have formed a complex with PXR. The identities and amounts of these protein bands were subsequently determined by Edman sequencing, which revealed less than one-to-one stoichiometry of SRC-1p to PXR. In addition, the purified protein sample appeared to be unstable and prone to precipitation at 4°C even at low concentrations (0.2 mg/ml), resulting in a low final yield (<5 mg/l of culture).
|
Co-expression of PXR and SRC1p using bicistronic constructs
Based on structural and biophysical analysis (Watkins et al., 2003a), the stability of PXR requires stoichiometric formation of a PXR/SRC-1p heterodimer. Bicistronic expression constructs may allow translation of the two proteins from the same transcript, at the same cellular location and potentially in an equimolar ratio. The first construct, D14-PXR-RBS-SRC, had PXR encoded by the first cistron followed by SRC-1p in the second cistron. PXR/SRC-1p constituted approximately 4.5% of the total cytosolic proteins, purification yielded 30 mg/l of culture (Fig. 2, lane 3). The purified sample (Fig. 2, lane 4) contained an approximate equimolar molar ratio of the two polypeptides based on quantitative Edman sequencing. The molecular mass of the two proteins was determined to be 36.3 and 9.3 kDa, respectively, using MALDI-MS, in good agreement with calculated values.
The purified PXR/SRC-1p complex was evaluated in binding studies using TdCD (Fig. 3A and B). A single unfolding transition was observed with a melting temperature (Tm) of 41.5°C for the complex in the absence of ligand (Table II). The Tm increased to 45.5 and 53.2°C in the presence of 10-fold excess of SRC-1p or SR-12813, respectively. In the presence of both SRC-1p and SR-12813, the Tm increased to 56.3°C, confirming an additional effect reported previously (Watkins et al., 2003a). Under our experimental conditions, the bicistron-derived protein complex appeared to be more stable than that produced using the dual plasmid system, although precipitation was still evident during purification and storage, resulting in >50% loss of total protein, and a decrease in the SRC-1p:PXR ratio.
|
) + SRC-1p; (x) + SR-12813; (o) + SRC-1p + SR-12813.
|
However, the second bicistronic construct, D14-SRC1p-RBS-PXR, which had a reversed cistron order, yielded only the His-tagged SRC-1p after the Ni-NTA step (Fig. 2, lane 6), and was therefore not investigated further.
Design and preparation of the tethered PXR-L10-SRC1p protein
Molecular modeling. In the PXR/SRC-1p crystal structure [PDB code 1NRL
[PDB]
(Watkins et al., 2003a)], PXR Ser-434, SRC1p 676-CPSSHS-681 and 697-GSPS-700 were disordered (Watkins et al., 2003a). The crystallographically observed C-terminus of PXR is near the N-terminus of SRC-1p, presenting a possibility for a covalent linker. It was therefore reasoned that a PXR/SRC-1p linked polypeptide may ensure an intramolecular, and thus concentration-independent, interaction. The distance from PXR Gly-433 to SRC-1p Ser-682 is 18 Å. Spanning this distance would require a minimum of five additional amino acid residues. In our design, SRC-1 residues Cys-676 and Pro-677 were removed to avoid potential conformational disruption due to potential disulfide bond formation and/or rigidity of the proline residue. The remaining disordered residues (i.e. PXR Ser-434 and SRC-1p S678SHS) were considered parts of the linkers; therefore, the single chain PXR/SRC-1p proteins with linkers of 8, 10 and 13 amino acid residues, respectively, were designed to encode the following sequences (those derived from SRC-1p are italicized):
| |||||||||||||||
In light of similar behavior exhibited by constructs containing the three linkers, the remainder of the analysis will focus on the construct with a decamer linker.
Expression of PXR-L10-SRC1p constituted approximately 5% of the soluble intracellular protein with a final purified yield of 70 mg/l of culture. The tethered protein eluted as an apparent dimer (MWapp=80 kDa) on a Superdex 200 gel filtration column and judged to be >95% pure based on SDS–PAGE (Fig. 4A and B). The molecular mass was 39 128 ± 2 Da as determined by ES-MS, in accordance with the calculated value (39 128 Da). The melting temperature (Tm) was determined to be 57.6°C (Fig. 4C, Table II), significantly higher than the recombinant PXR/SRC-1p complexes derived from the bicistronic expression system (Tm=45.5°C, Fig. 3B, Table II). A 6°C increase in Tm was observed in response to additional 10-fold of SR-12813 to the tethered construct; addition of exogenous SRC-1p to either the ligand-bound or ligand-free form of the tethered construct did not alter the Tm value (Fig. 4C, Table II).
|
Evaluation of compound binding affinity to PXR-L10-SRC1p
In the previously solved binary complex PXR/SR-12813, the ligand was present simultaneously in three different orientations (Watkins et al., 2002); adding SRC-1p restricted the ligand to a single orientation in the ternary complex PXR/SRC-1p/SR-12813, suggesting that SRC-1p may have restricted ‘breathing’ of PXR and thus ‘trapped’ the ligand in the productive orientation (Watkins et al., 2003a). Since the PXR/SRC-1p complex derived from both the dual plasmid and bicistronic expression systems requires additional SRC-1p for full complexation, the unsaturated PXR fraction in these protein samples may lead to underestimation of ligand affinity. Since the Tm value of the tethered protein was not affected by additional SRC-1 peptide, the PXR-LBD domain must have been effectively saturated with the linked peptide, rendering it constitutively ‘active’, and therefore a more physiologically relevant reagent for evaluating the affinity of PXR ligands in vitro.
To test this hypothesis, the thermal stability of the tethered construct PXR-L10-SRC1p in the presence of several well-known PXR binders was measured using circular dichroism (Table III). Hyperforin is an active ingredient in St John’s Wort, a medicine known to accelerate the metabolism of many prescription medicines (Mannel, 2004). Upon binding, it increased the Tm value 6.8°C compared with ligand-free PXR-L10-SRC1p, the most dramatic increase observed in this study, followed by clotrimazole (5.6°C), troleandomycin (5.4°C), troglitazone (3.4°C) and sulfinpyrazole (2.3°C). PCN has been shown to be a strong inducer of mouse PXR but a rather weak one for the human counterpart (LeCluyse, 2001). In our study, it showed only a marginal increase in Tm (0.6°C).
|
The compounds’ binding affinity was compared with CYP 3A4 mRNA induction employing human hepatocytes. While the rank order of compound binding and mRNA induction was not identical, the compounds which showed positive binding in TdCD also demonstrated significant induction, ranging from 10- to 25-fold. On the other hand, PCN showed no significant binding or mRNA induction. These data suggest that the TdCD assay may provide a rapid format to determine binding affinity of large numbers of compounds prior to confirmation in a more physiologically relevant system. Nevertheless, cell-based induction data are necessary in order to differentiate agonists from antagonists and to gauge the in vivo physiological response, which requires both ligand binding and downstream transcription.
Crystallization and structure solution
Crystals have been grown in which PXR-LBD is tethered to SRC-1p via flexible linkers of 8, 10 or 13 residues. Crystals appeared in as little as 24 h as thick rods, frequently with depletion-zones at their ends giving a crown-shaped appearance to each end of the crystal. Surprisingly, the diffraction quality of these crystals was indistinguishable from those which lacked this intergrown appearance. Crystals continued to grow to a maximum final size of 400 µm long by 150 µm across the narrow portion. The remainder of the structural analysis focuses on the tethered construct containing a decamer linker. Crystals of PXR-L10-SRC1p contained two monomers per asymmetric unit and belonged to the same orthorhombic crystal form previously described for the ternary complex of PXR-LBD with SR-12813 and exogenous SRC-1 peptide (Watkins et al., 2003a). Solution of the crystal structure using molecular replacement was straightforward. Final crystallographic data collection and refinement statistics are summarized in Table IV.
|
Structural analysis
Refined structures of the ligand-free tethered construct, PXR-L10-SRC1p apoprotein, and its complex with SR-12813 were compared pairwise with previously reported apoprotein and ligand complex structures. The root-mean-square deviation (rmsd) over all PXR-LBD C atoms or the 28 residues which contribute to the ligand binding site [as defined by Watkins et al.; (Watkins et al., 2001)] were calculated to be less than 1 Å in all cases (data not shown). When only the side-chain atoms of those 28 residues are considered, the rmsd values show a marked increase, but no consistent difference between the various apoprotein and ligand-bound forms (data not shown).
An overall structural alignment between the previously reported PXR-LBD in complex with both SRC-1p and SR-12813, and the tethered construct complexed with the same ligand, yields an rmsd of 0.29 Å over 259 matching C atoms. There are no significant changes at or near the SRC-1p binding site, including the AF-2 region. The most significant changes between the two structures are in some of the side chains of residues which line the ligand-binding site. Specifically, the side chains of His-407, Met-243, Phe-420, Phe-429 and Leu-206 to Ser-208 on helix 2 undergo maximal shifts of 0.5 to 1.5 Å while the corresponding main-chain atoms show a maximal shift of about 0.5 Å.
The crystal structure of the PXR-L10-SRC1p apoprotein presented here represents the first visualization of a PXR-LBD lacking a ligand while the SRC-1 coactivator peptide is bound. Interpretable electron density at the C-terminus of PXR-LBD ends at residue 432 (monomer A) or 431 (monomer B). Residues corresponding to 682 to 696 (monomer A) or 682 to 698 (monomer B) of the tethered SRC1 peptide are observed (Fig. 5). The linker residues are sufficiently disordered that they are not interpretable in difference Fourier electron density maps. The ligand-binding cavity shape and volumes were compared among the tethered PXR-LBD structures, the published coactivator-free apoprotein, and the complex with SRC-1p and SR-12813. As shown in Fig. 6B, the shape of the ligand-binding cavity in the tethered structures more closely resemble that of the previously described activated complex with SR-12813, even when the ligand is not present.
|
. The figure is presented in wall-eyed stereo.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
PXR-mediated induction of metabolic enzymes is the body’s primary defense system in response to a myriad of ingested, inhaled or otherwise absorbed xenobiotics. It is a prominent regulator of CYP3A, a major constituent in drug metabolism. Many chemicals, including drugs currently in use, stimulate PXR-mediated CYP induction (Kliewer et al., 2002
). Studies have identified differences in the activation profiles of PXR across species in response to various xenobiotics (Tirona et al., 2004).
A typical transcription factor, PXR is transiently expressed and tightly regulated to allow precise coupling of ligand binding and downstream gene expression (Honkakoski et al., 2003) (Watkins et al., 2001). While PXR recognizes a plethora of chemical entities, it also favors certain groups of ligands. In addition, the plasticity of the ligand-binding site is restricted upon binding to SRC-1p (Watkins et al., 2003a), suggesting that coactivator peptide binding is correlated with the productive conformation of PXR-LBD.
Watkins et al. have employed a dual plasmid system to generate a PXR-LBD/SRC1 protein complex (Watkins et al., 2001). However, in the current study, we were unable to achieve stochiometric expression of the proteins using the same strategy. The imbalance was likely due to the different in vivo copy numbers determined by the two replicons in the two plasmids (i.e. colE 1 (in pDEST14): p15A(in pACYC184)=4:1) (Martinez et al., 1988; Jobling and Holmes, 1990). As a result, the majority of the purified PXR-LBD was uncomplexed with the coactivator and therefore susceptible to aggregation and precipitation.
In prokaryotic cells, a polycistronic configuration, in which two or more open reading frames are regulated under a single strong promoter, can be used as an effective strategy to express multiple polypeptides in equimolar ratios by generating the same numbers of transcripts in vivo (Nagami et al., 1988; Tan, 2001). In the current study, an oligodeoxynucleotide spacer containing a stop codon and a T7 gene 10 RBS was placed between the two open reading frames to avoid a fusion protein, and the DNA fragment was regulated by a strong T7 promoter in vector pDEST14 (Invitrogen, Carlsbad, CA, USA). While the SRC-RBS-PXR configuration did not lead to detectable expression, PXR-RBS-SRC not only resulted in co-purification of the two polypeptides in an equimolar ratio (only PXR-LBD contained an N-terminal hexahistidine tag), but also 5-fold higher expression level and increased protein stability. It was not clear why the order of the open reading frames had such a dramatic impact on the expression levels. One possible explanation was that the tempo of transcription and/or translation of the two proteins may have to synchronize precisely to ensure efficiency of complexation. Similar discrepancies also have been observed for the and β subunits of human farnesyltransferase in a bicistronic system (Wu et al., 1999). However, both of the expressed proteins (especially SRC-1p) still precipitated (albeit at approximately one-third the rate of the dual-plasmid derived counterparts) during the process of purification, compromising the final yield of purified protein. This observation suggested that the intrinsic affinity between the proteins may not be high enough to eliminate oscillation of the PXR between ‘on’ (SRC-1p bound, stable) and ‘off’ (SRC-1p not bound, unstable) modes. This hypothesis was further supported by the finding that additional SRC-1p significantly diminished precipitation and increased the thermal stability of the purified complex (Tm was 41.5 and 45.5°C, respectively, in the absence and presence of 10-fold molar excess of SRC-1p; see Table II).
A main purpose of this study was to generate a homogenous and active PXR-LBD/SRC-1p complex which could be used to rapidly screen many chemical compounds for a possible PXR liability. Structural analysis indicated that it may be possible to covalently connect the two proteins into a fusion polypeptide, creating an intramolecular complex that may be more stable than the intermolecular one. Similar approaches have been utilized to create a single-chain HCV protease-NS4A cofactor (Taremi et al., 1998) and HIV-1 P66-P51 fusion reverse transcriptase (Zuniga et al., 2004). All three designed constructs (PXR-L8, 10, 13-SRCp) led to robust expression and further analysis of PXR-L10-SRC1p revealed increased thermal stability relative to the PXR/SRC-1p heterodimer. Adding exogenous SRC-1p did not further increase the Tm value, suggesting that fused SRC-1p has saturated its binding site on PXR, locking the ‘on’ mode.
Circular dichroism spectra of the PXR/SRC-1p complex and the tethered proteins were superimposable, confirming that this novel engineering endeavor has preserved the secondary and tertiary structures of PXR. SR-12813 and several other drugs increased the thermal stability of these proteins. The ligands which stabilized the tethered construct also showed positive responses in the induction of CYP 3A4 mRNA in human hepatocytes, though the rank order was not identical. This observation suggests that these single-chain PXR/SRC-1p proteins are in the activated conformation and thus suitable for rapid evaluation to predict potential drug–drug interactions.
Previous structural analyses focused on the effects of ligand binding to the apoprotein (Watkins, Science, 2001) and coactivator peptide binding to the ligand-bound form (Watkins et al., 2003b). The structure of unliganded PXR-LBD-L10-SRC1p presented here complements those previous studies by providing a glimpse of the effects of coactivator binding in the absence of a ligand. This structure allows the dissociation of conformational changes due solely to ligand binding from those due to binding of coactivator peptide, and supports the observation that the tethered construct remains in an activated conformation even in the absence of a bound ligand. Interestingly, the tethered protein appeared to be a homodimer in solution as well as in the structures, in good agreement with the tryptophan-zipper mediated homodimer model recently proposed (Noble et al., 2006). This observation further supports that this newly designed recombinant PXR is functionally similar to the PXR/SRC-1p dimer, and is easier to produce for large scale binding screens and structural determinations.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Funding for this work was provided by the Schering-Plough Research Institute, Kenilworth, NJ, USA.
|
Footnotes |
|---|
Edited by Bengt Mannervik
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
The authors thank Rumin Zhang for peptide synthesis, Yan-Hui Liu for MALDI-TOF and electrospray mass spectrometry, and Charles McNemar for peptide sequencing.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Blumberg B., et al. Genes Dev. (1998) 12:3195–3205.
Bricogne G. Acta Crystallogr. D (1993) 49:37–60.[CrossRef][Medline]
Bricogne G. Macromolecular Crystallography.—Carter C.W., Sweet R.M., eds. (1997) vol. 276. Academic Press. 361–423.[CrossRef]
Bricogne G., Irwin J.J. (1996) Macromolecular Refinement, Proceedings of the CCP4 Study Weekend, SERC Daresbury Laboratory. Warrington, England.
Buczek O., et al. Protein Sci. (2002) 11:924–932.[CrossRef][Web of Science][Medline]
Collaborative Computational Project, Number 4. Acta Crystallogr. D (1994) 50:760–763.[CrossRef][Medline]
Cui X., et al. J. Pharmacol. Toxicol. Methods (2005) 52:234–243.[CrossRef][Medline]
Darimont B.D., et al. Development (1998) 12:3343–3356.
DeLano W.L. The PyMOL Molecular Graphics System (2002) Alto CA, USA: DeLano Scientific, Palo.
Gampe R.T. Jr, et al. Mol. Cell (2000) 5:545–555.[CrossRef][Web of Science][Medline]
Handschin C., Meyer U.A. Pharmacol. Rev. (2003) 55:649–673.
Honkakoski P. Curr. Pharmacogenomics (2003) 1:75–85.[CrossRef]
Honkakoski P., et al. Ann. Med. (2003) 35:172–182.[CrossRef][Web of Science][Medline]
Jobling M.G., Holmes R.K. Nucleic Acids Res. (1990) 18:5315–5316.
Jones T.A., et al. Acta Crystallogr. A (1991) 47:110–119.[CrossRef]
Jones S.A., et al. Methods Enzymol. (2002) 357:161–170.[Web of Science][Medline]
Kleywegt G.J., Jones T.A. Acta Crystallogr. D (1994) 50:178–185.[CrossRef][Medline]
Kliewer S.A., et al. Cell (1998) 92:73–82.[CrossRef][Web of Science][Medline]
Kliewer S.A., et al. Endocrinol. Rev. (2002) 23:687–702.
LeCluyse E.L. Chem. Biol. Interact. (2001) 134:283–289.[CrossRef][Web of Science][Medline]
Mannel M. Drug Saf. (2004) 27:773–797.[CrossRef][Web of Science][Medline]
Martinez E., et al. Gene (1988) 68:159–162.[CrossRef][Web of Science][Medline]
Moore J.T., Kliewer S.A. Toxicology (2000) 153:1–10.[CrossRef][Web of Science][Medline]
Moras D., Gronemeyer H. Curr. Opin. Cell Biol. (1998) 10:384–391.[CrossRef][Web of Science][Medline]
Nagami Y., et al. Gene (1988) 69:59–69.[CrossRef][Web of Science][Medline]
Noble S.M., et al. Biochemistry (2006) 45:8579–8589.[CrossRef][Web of Science][Medline]
Nolte R.T., et al. Nature (1998) 395:137–143.[CrossRef][Medline]
Raucy J., et al. J. Pharmacol. Exp. Ther. (2002) 303:412–423.
Roversi P., et al. Acta Crystallogr. D (2000) 56:1316–1323.[CrossRef][Medline]
Tan S. Protein Expr. Purif. (2001) 21:224–234.[CrossRef][Web of Science][Medline]
Taremi S.S., et al. Protein Sci. (1998) 7:2143–2149.[Web of Science][Medline]
Tirona R.G., et al. Mol. Pharmacol. (2004) 65:36–44.
Tronrud D.E., Ten Eyck L.F., Matthews B.W. Acta Crystallogr. A (1987) 43:489–501.[CrossRef]
Wang H., LeCluyse E.L. Clin. Pharmacokinet. (2003) 42:1331–1357.[CrossRef][Web of Science][Medline]
Watkins R.E., et al. Science (2001) 292:2329–2333.
Watkins R.E., et al. Curr. Opin. Drug Discov. Dev. (2002) 5:150–158.[Web of Science][Medline]
Watkins R.E., et al. J. Mol. Biol. (2003) a 331:815–828.[CrossRef][Web of Science][Medline]
Watkins R.E., et al. Biochemistry (2003) b 42:1430–1438.[CrossRef][Web of Science][Medline]
Weatherman R.V., et al. Ann. Rev. Biochem. (1999) 68:559–581.[CrossRef][Web of Science][Medline]
Willson T.M., Kliewer S.A. Nat. Rev. Drug Discov. (2002) 1:259–266.[CrossRef][Web of Science][Medline]
Wu Z., et al. Protein Eng. (1999) 12:341–348.
Xu H.E., et al. Proc. Natl Acad. Sci. USA (2001) 98:13919–13924.
Zuniga R., et al. Protein Eng. Des. Sel. (2004) 17:581–587.
Received January 2, 2008; revised March 14, 2008; accepted March 15, 2008.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||