Effects of Co-Expression of Liver X Receptor β-Ligand Binding Domain with its Partner, Retinoid X Receptor α-Ligand Binding Domain, on their Solubility and Biological Activity in Escherichia coli

Abstract

In this presentation, I describe the expression and purification of the recombinant liver X receptor β-ligand binding domain proteins in E. coli using a commercially available double cistronic vector, pACYCDuet-1, to express the receptor heterodimer in a single cell as the soluble form. I describe here the expression and characterization of a biologically active heterodimer composed of the liver X receptor β-ligand binding domain and retinoid X receptor α-ligand binding domain. Although many of these proteins were previously seen to be produced in E. coli as insoluble aggregates or "inclusion bodies", I show here that as a form of heterodimer they can be made in soluble forms that are biologically active. This suggests that co-expression of the liver X receptor β-ligand binding domain with its binding partner improves the solubility of the complex and probably assists in their correct folding, thereby functioning as a type of molecular chaperone.

Introduction

Members of the nuclear receptor superfamily including the receptors for retinoic acid, vitamin D, thyroid hormone, and steroid hormones are hormone-activated transcription factors that have important roles in many physiological processes [12,13]. The hormone nuclear receptor superfamily includes different proteins responsible for the direct link between transcriptional regulation and physiological responses. A subgroup of receptors within the superfamily, including the retinoic acid receptor (RAR), thyroid hormone receptor (TR), peroxisome proliferator activated receptor (PPAR), and many orphan receptors, bind to specific DNA sequences known as hormone response elements as heterodimers with the retinoid X receptor (RXR) and share a conserved structural and functional framework [12,14]. The nuclear receptors have a modular structure composed of three independent conserved domains; the N-terminal AF-1, the highly conserved DNA binding domain, and finally the ligand binding and dimerization domain with liganddependent AF-2 region [2]. Liver X receptor (LXR, consisting of two members, LXR α and LXR β, are encoded by Nr1b3 and Nr1b2, respectively) is an orphan receptor that heterodimerizes with RXR, binds to a specific response element called the LXRE, and has the potential to modulate physiological responses [2]. Both LXRs are widely expressed in multiple tissues and cell types, but LXR α is most highly expressed in the liver and intestine [28].

The natural ligands for the LXRs consist of a select group of oxysterols derived from tissue-specific cholesterol metabolism in the liver, brain, and gonads [21]. Although their presumed natural ligand is not cholesterol itself, recent studies suggest that the LXRs serve as one of the body’s key sensors of dietary cholesterol and therefore regulate essential pathways in cholesterol homeostasis. Activated in response to increased levels of intracellular cholesterol, LXRs act to induce specific gene expression programs to control various aspects of cholesterol homeostasis [9,21]. In addition, LXR signaling negatively regulates inflammation by repressing the expression of genes involved in the inflammatory response [10,19].

The three-dimensional structure of the LXR β-LBDs, in complex with the synthetic agonist T-0901317, has been determined [7,24,27,29]. Analysis of the structure provides evidence that the heterodimeric partners communicate with each other, presumably through their dimer interface and also potentially through their associated co-factors [6]. Physiologically, activation of LXRs drives the expression of genes involved in reverse cholesterol transport, biliary excretion, and intestinal absorption [21], and could lead to an increase in HDL particle numbers. In mice, treatment with synthetic agonists results in increased HDL levels and decreased atherosclerotic lesions [11,25]. Thus, LXRs may be promising therapeutic targets for treating dyslipidemia and atherosclerosis. LXR β is likely to be a good target even though it is ubiquitously expressed; for example, LXR β agonists would be useful in non-liver tissues such as macrophages and adipose tissues.

However, the development of LXR ligands as therapeutic agents has been hampered by the increase in hepatic triglyceride production by the dual LXR agonists [15,26]. LXR α is the dominant subtype in the liver, intestine, kidney, spleen, and adipose tissue, whereas LXR β is ubiquitously expressed at very low levels. In fact, LXR α leads to stimulation of fatty acid synthesis pathways and increased triglyceride levels. Therefore, an LXR β-selective agonist might retain efficacy without deleteriously increasing hepatic lipogenesis [20]. Hence, the identification of LXR β-selective agonists may be of significant therapeutic value.

Our aim was to express and purify the recombinant LXR β-LBD proteins in E. coli using a commercially available double cistronic vector, pACYCDuet-1, to express the receptor heterodimer in a single cell. I describe here the expression and characterization of the LXR β-LBD and its heterodimeric partner, RXR α-LBD. When expressed alone, the LXR β-LBD was found to be largely insoluble and the formation of heterodimeric complexes was not sufficient. However, by using a double cistronic expression system, I was able to obtain high levels of soluble, tightly associated heterodimers. This suggests that co-expression of LXR β-LBD with its binding partner RXR α-LBD improves the solubility of the complex and probably assists in their correct folding, thereby functioning as a type of molecular chaperone.

 

Materials and Methods

Preparation of Expression Constructs

For the expression of (His)6 -tagged LXR β-LBD(205-461) and RXR α-LBD(225-462), cDNA fragments encoding the human LXR β-LBD and RXR α-LBD were obtained by RT-PCR and subcloned into pET-15b (driven by the PT7 promoter; Novagen, Darmstadt, Germany), pGEX-4T-1 (driven by the PTac promoter; GE Healthcare, Piscataway, NJ, USA), and pACYCDuet-1 (driven by the PT7 promoter; Novagen) vectors (Fig. 1). For the pACYCDuet-1 vector, (His)6-tagged LXR β-LBD(205-461) was inserted into the 1st MCS at the BamHI/HindIII site and RXR α-LBD(225-462) into the 2nd MCS at the EcoRV/XhoI site. All constructs were confirmed by DNA sequencing. The oligonucleotide primers used for PCR-based cloning are shown in Table 1. All molecular biology techniques for constructing the expression plasmids were performed using standard procedures.

Fig. 1.Schematic diagram of the expression constructs. (A and B) The single cistronic expression constructs in pET-15b and pGEX-4T-1, and western blot analysis for detecting the recombinant LXR β-LBD proteins using anti-(His)6 and anti-GST antibodies. Tac and T7 promoters (PT7 and PTac) are indicated by arrows. Repressor binding sites are represented by open boxes and E. coli ribosome binding sites by open circles. S: soluble fraction; P: insoluble particulate fraction. (C) The double cistronic expression construct in the pACYCDuet-1 expression vector. LXR β-LBD was fused to (His)6 at its N-terminal and subcloned into the 1st multiple cloning site, whereas RXR α-LBD was not fused to any tag and subcloned into the 2nd site.

Table 1.Underlined sequences represent the restriction enzyme recognition sites.

Protein Expression and Purification

The expression constructs for LXR β-LBD in pET-15b, pGEX4T-1, and pACYCDuet-1 were used to transform BL21 (DE3) Pro E. coli (Invitrogen, Carlsbad, CA, USA). Cultures were grown at 37℃ in LB medium containing antibiotics. These cultures were induced at an optical density of 0.8-1.0 (detected at 600 nm) using 0.4 mM isopropyl β-D-thiogalactopyranoside (IPTG) and grown for a further 5-6 h before harvesting by centrifugation. The collected cell pellets were resuspended in buffer A containing 50 mM Tris-HCl, pH 8.0, 100 mM KCl, and 1 mM EDTA for purifying glutathione S-transferase (GST) fusion protein or in buffer B containing 20 mM HEPES, pH 8.0, 100 mM HCl, and 20 mM imidazole for purifying the (His)6-tagged proteins. The cells were lysed by ultrasonication in the above-mentioned buffers containing Protease Inhibitor Cocktail (Roche Diagnostic, Manheim, Germany). After lysis by ultrasonication, the lysate was cleared by centrifugation and analyzed immediately or stored at -80℃. For purification, the lysates were loaded at 4℃ onto a Ni+2-nitrilotriacetate acid (NTA) agarose column (Probonds Resin, Invitrogen). The agarose columns were washed with either buffer A or buffer B. The proteins were eluted from the column in the same buffer containing the appropriate concentration of imidazole. The fractionation was performed with a Source 15Q anion-exchange column (GE Healthcare, Piscataway, NJ, USA, column volume (CV) ¼ 3.9 ml) with an NaCl gradient. Fifteen milliliters of the desalted lysate (15 mg of total protein) was loaded and eluted with three-step linear gradients (25–200 mM NaCl for 1.5 CV, 200–300 mM NaCl for 4 CV, and 300–700 mM for 2 CV) to evenly distribute the proteins into separate fractions. Fifteen 1.5 ml fractions were collected and used immediately or stored at -80℃. Chromatography w as performed w ith the AKTA FPLC system (GE Healthcare; Piscataway, NJ, USA).

Solubility Analysis

E. coli cell pellets were resuspended in 50 mM Tris-HCl, pH 8.0, 100 mM KCl, 1 mM EDTA, and Protease Inhibitor Cocktail, ultrasonicated, and then centrifuged at 3,000 ×g for 20 min to remove intact cells and cell debris. The supernatants were then further centrifuged at 10,000 ×g for 20 min to precipitate the insoluble proteins. Soluble proteins were purified from the supernatant.

Ligand Binding Using Fluorescence Polarization Assays

Fluorescence polarization assays were performed using a Victor apparatus (Perkin-Elmer Life and Analysis Sciences, Boston, MA, USA) according to the method described by Dagher et al. [5] and Lévy-Bimbot et al. [18]. Assays were conducted in 96-well black polystyrene plates. The effects of T-0901317 on the interaction of the co-activator peptide (SRC-1a, CPSSHSSLTERHKILHRLLQEGSPS; Cosmo Genetech Co., Seoul, Korea) and the acceptor APC with LXR β-LBD/RXR α-LBD were determined by fluorescence polarization assays in the assay buffer. Varied concentrations of LXR β-LBD/RXR α-LBD in the presence or absence of T-0901317 were incubated at room temperature with a europirim-labeled peptide. The polarization degrees (FP) were measured with an excitation wavelength of 615 nm and an emission wavelength of 665 nm. The apparent dissociation constant values were determined by the binding curves derived from a non-linear least-squares fit of the data for a sample 1:1 interaction.

Western Blot Analysis and Antibodies

Equal amounts of protein were loaded and resolved by 12% SDS-PAGE. Resolved proteins were transferred to an NC membrane (S&S, Dassel, Germany), blocked with 5% non-fat dry milk in TBS, and probed with specific antibodies against (His)6-tag, GST, RXR α, and LXR β (Cell Signaling Technology, Beverly, MA, USA), followed by the secondary antibodies coupled to horseradish peroxidase (Bio-Rad, Hercules, CA, USA). The immunoreactive proteins on the membrane were detected by chemiluminescence using the Western Blotting Detection Reagent Kit (AbSignal, AbClon Co., Seoul, Korea).

 

Results

Expression of LXR β-LBD Proteins from PT7 and PTac Promoters as Single Cistronic Plasmid Construction

Nr1b2 (LXR β-LBD) was cloned into the pET-15b expression vector to produce LBD tagged at the N-terminus with six histidine residues, and also into the pGEX-4T-1 vector so as to express a GST-tagged protein. The clones were used to transform E. coli BL21 (DE3) pro for an initial expression test. E. coli cells, induced to express recombinant protein by IPTG, formed both insoluble particulate and soluble fractions. As shown in the right panels of Figs. 1A and 1B and Fig. 2B, LXR β-LBD(205-461) was found to be insoluble at a high level using both single expression constructs. The insoluble LXR β-LBD(205-461) could be detected by anti-(His)6 antibodies by western blot analysis (Fig. 2A). In addition, the expressed RXR α-LBD could not be pulled down by GST LXR β-LBD on agarose beads (data not shown). These results demonstrate that the single citronic pET and pGEX vectors driven by T7 and Tac promoters failed to support the correct folding of the proteins under our experimental conditions.

Fig. 2.Expression of recombinant LXR β-LBD using pET-15b expression vector. The LXR β-LBD(205-461) gene was inserted into pET-15b containing a hexahistidine sequence to the construct. Soluble (S) and insoluble (P) proteins were separated on an SDS-PAGE gel, subjected to immunoblot analysis with anti-(His)6 antibody (A), and stained with Coomassie Brilliant Blue R-250 (B). The arrow indicates the expressed proteins.

Double Cistronic Plasmid Construction, Expression, and Purification

Several lines of evidence suggest that the heterodimeric partners communicate with each other, presumably through their dimer interface and/or their associated co-factors [17]. Thus, I used a commercially available double cistronic vector, pACYCDuet-1, that directs expression of the heterodimer in a single cell. LXR β-LBD(205-461) was cloned into the pACYCDuet-1 expression vector, which is designed to co-express two target genes (Fig. 1C). The vector contains two multiple cloning sites, each of which is driven by a T7 promoter (PT7) and preceded by a lac operator and ribosome binding site (Rbs). In the cistronic construct, LXR β-LBD(205-461) tagged with (His)6 at its N-terminus is placed upstream of RXR α-LBD(225-462), its expression partner in the so-called pACYCDuet/HN-LXR β-LBD/RXR α-LBD vector. The double cistronic vectors gave high yields of both soluble proteins (Fig. 3). The E. coli BL21 (DE3) cells transformed with the pACYCDuet/HN-LXR β-LBD/RXR α-LBD vector (Fig. 1C) were grown at 37℃ and induced with 0.4 mM IPTG at either 20℃ or 30℃ for 5-6 h. The cells were harvested and lysed by ultrasonication, and the soluble (S) and insoluble (P) fractions were collected and analyzed by SDS-PAGE (Fig. 3). Although the two partners (LXR β-LBD and RXR α-LBD) were expressed primarily as insoluble proteins at both temperatures, the recombinant proteins were detected as soluble proteins (Fig. 3).

Fig. 3.SDS-PAGE analysis of recombinant fusion protein expression (LXR β-LBD(205-461)) using a double cistronic expression system with RXR α-LBD(225-462). Soluble (S) and insoluble (P) recombinant LXR β-LBD(205-461) fusion proteins induced or uninduced at either 20℃ or 30℃ and resolved by SDS-PAGE and stained with Coomassie Blue R-250. The arrow indicates the position of the recombinant LXR β-LBD(205-461) fusion protein with its binding partner, RXR α-LBD(225-462). Lane M represents the calibrated molecular weight markers.

The LXR β-LBD was found to be soluble and expressed at relatively high levels by co-expressing its binding partner, RXR α-LBD, using the double cistronic expression vector (Figs. 1C and 3). The transformed E. coli cells were harvested after induction with 0.4 mM IPTG at 30℃. The bacterial lysate was obtained using an ultrasonicator, and, after clarification by centrifugation, was loaded onto a Ni+2-NTA agarose column. The column resin was then washed until no protein was found in the eluate. The bound proteins were eluted stepwise using 40, 100, and 300 mM of imidazole, purified, resolved by SDS-PAGE, and visualized with Coomassie Brilliant Blue R-250. In this way, the highly pure (at least 95%) heterodimers of LXR β-LBD and RXR α-LBD were obtained (Fig. 5). To decrease the complexity of the E. coli protein for target identification, the soluble fraction of an E. coli lysate was fractionated by anion-exchange chromatography. The salt gradient was chosen after multiple trials to maximize the distribution of proteins into different fractions. The chromatogram (absorbance at 280 nm) and the salt profile (conductivity of the elution buffer) are shown in Fig. 5A. The eluant from 3 to 20 ml was separated into fifteen 1.5 ml fractions. Proteins that were eluted at the washing step were not used in this study. SDS-PAGE of the fractions showed that proteins in the cell lysate were evenly fractionated (Fig. 5B). Most fractions contained 30 to 50 easily identifiable protein bands on the SDS-PAGE gels.

Fig. 4.Analysis of heterodimer formation for the recombinant LXR β-LBD fusion protein with its binding partner, RXR α-LBD, using a double cistronic expression vector. Cell lysates from the induced [IPTG (+)] or uninduced [IPTG (-)] E. coli cells at either 20 ℃ or 30 ℃ were incubated with either anti-(His)6 antibody (A) or anti-LXR antibody (B) and protein A agarose. The affinity-purified proteins were resolved by SDS-PAGE and subjected to immunoblot analysis with antibodies against the RXR α protein.

Fig. 5.Purification of the recombinant protein using a double cistronic expression vector, pACYCDuet-1. (A) An E. coli lysate was fractionated by using anion-exchange chromatography. The chromatogram shows the elution monitored by absorbance at 280 nm (A280; solid line) and the conductivity (dashed line). The location of each fraction is indicated by a vertical tick on the x-axis. (B) The protein contents of the fractions were checked by SDS-PAGE. The molecular weight markers (M) and the total lysate (L) were loaded as references.

Analysis of Ligand Binding

In this study, I used T-0901317 as an agonist to evaluate the biological activity of the recombinant LXR β-LBD/RXR α-LBD heterodimer.

The effects of T0901317 on the interaction of the co-activator peptide (SRC-1a, CPSSHSSLTERHKILHRLLQEGSPS) and the acceptor APC with the recombinant LXR β-LBD/RXR γ-LBD protein were determined by fluorescence polarization assays. Varied concentrations of the purified recombinant LXR β-LBD/RXR α-LBD in the presence or absence of T-0901317 were incubated at room temperature with a europirim-labeled peptide. The polarization degrees were measured with an excitation wavelength of 615 nm and an emission wavelength of 665 nm. Fig. 6 shows a typical dose-response curve obtained by using increasing concentrations of the purified recombinant LXR β-LBD/RXR α-LBD and measuring FI as an index of fluorescent peptide binding to LBD. This type of experiment was repeated a number of times during the course of this study, allowing us to calculate an ED50 of 719 nM. Because our assays were performed using only the purified recombinant LXR β-LBD/RXR α-LBD protein, it is likely that the true value could somewhat lower. Thus, a protein expression system using the dual cistronic pACYCDuet-1 vector with IPTG induction was found to be suitable for producing soluble heterodimers consisting of LXR β-LBD(205-461) and RXR α-LBD(225-462) and this should be helpful for future biological and structural analyses.

Fig. 6.Ligand-binding studies of the purified recombinant protein using a fluorescence polarization assay. To detect direct binding of probes to synthesized and purified heterodimers of LXR β-LBD and RXR α-LBD by the co-activator peptide (SRC-1a, CPSSHSSLTERHKILHRLLQEGSPS; Cosmo Genetech Co., Seoul, Korea) and the acceptor APC, a fluorescence polarization assay was performed as described in Materials and Methods. Each datum point was measured in triplicate. The polarization degree (FP) was plotted against the recombinant protein concentration, and non-linear regression analysis was performed to determine the ED50.

 

Discussion

High expression of functional proteins in E. coli, especially those from eukaryotic origin, has often been challenging as the proteins tend to aggregate in inclusion bodies within the bacteria. Recently, the three-dimensional structure of LXR β-LBDs was determined, although expression of LXR β-LBD in E. coli yielded predominantly insoluble proteins [7,24,27]. The soluble fraction was unstable and could not be suitably concentrated for protein crystallization. To obtain the active protein, the aggregated protein was solubilized with detergent or denaturing agents and re-folded to recover the protein activity. However, frequently re-folded proteins may not recover their biological activity. Despite some limitations, E. coli remains the most efficient and widely used host for recombinant protein production [23]. The main reasons for using E. coli as a host for expressing recombinant proteins include its well-defined genetics, high transformation efficiency, simple cultivation, short doubling time (rapidity), and inexpensiveness. In fact, preparations to enrich a specific protein are rarely easily obtained from their natural host cells. Hence, recombinant protein production is frequently the sole applicable procedure. With the advent of the post-genomic era has come the need to express in E. coli a growing number of genes originating from different organisms. Recombinant fusion partners and mutant strains have advanced the possibilities with E. coli.

To explore the interaction between LXR β-LBD and RXR α-LBD, the (His)6-tagged LXR β-LBD proteins (and any interacting proteins) were recovered by immunoprecipitation using antibodies against LXR and (His)6, and the LXR β-LBD immunoprecipitate was subjected to SDS-PAGE and analyzed for the presence of RXR α-LBD. As shown in Figs. 4A and 4B, the recombinant RXR α-LBD protein was observed in both the immunoprecipitates collected with anti-(His)6 and anti-LXR antibodies. This co-immunoprecipiation approach has the advantage of being able to identify the direct interaction with LXR β-LBD that is relevant in this biological system. In addition, this dimer was found to be stable under other conditions (data not shown). From these experiments, I have demonstrated that the expression of LXR β-LBD using the single cistronic expression vectors are in some ways defective for proper folding and biological activities. However, the bacterially produced heterodimer using the double cistronic expression vector is soluble and the two dimmers are not easily dissociated. It is clear that co-expression increases the solubility of LXR β-LBD as well as ensuring its proper folding. This phenomenon might be due to the masking of any hydrophobic patches that can cause LXR β-LBD to form aggregates [22].

Our method was used to produce a soluble recombiant protein by using a double cistronic E. coli expression system. Because the soluble fraction can be directly applied to Ni-NTA columns for purification, there are advantages. In addition, use of the pACYCDuet-1 vector further simplifies downstream purification processes, including cleavage and elution. Thus, our approach may be helpful for increasing the solubility and yield of other proteins of interest.

The actions of lipophilic hormones, including steroids, retinoids, thyroid hormone, and vitamin D3, are mediated through a conserved superfamily of nuclear receptor proteins that function as ligand-regulated, DNA-binding transcriptional activators [16]. LXRs were initially described as orphan receptors; later, monooxygenated cholesterol derivatives, oxysterols, were identified as their natural ligands [8,9]. Recently, two nonsteroid synthetic LXR agonists, T-0901316 and GW3965, have been widely used as nonsteroidal chemical tools to explore the biology of LXRs. T-0901317 activates both LXR α and LXR β, whereas GW3965 has a greater affinity for LXR β than for LXR α [3,4,8].

I have successfully expressed and purified the LXR β-LBD proteins as a heterodimer with RXR α-LBD in E. coli using a double cistronic expression vector system. The recombinant LXR β-LBDs are substantially soluble and could be useful for future biological, biophysical, and structural analyses of nuclear receptor complexes. Our approach could be useful for achieving high expression levels of other nuclear receptors and may be applied as well to other classes of heterodimeric protein partners. Finally, our results prove that the recombinant LXR β-LBD could be a promising target for the development of molecular ligands with improved therapeutic windows.