Protein Expression and PuriWcation 42 (2005) 137–145 www.elsevier.com/locate/yprep
Expression, puriWcation, and characterization of recombinant human Xotillin-1 in Escherichia coli Yu Ding, Ming Jiang, Weihua Jiang, Yang Su, Hanqing Zhou, Xiaojian Hu, Zhihong Zhang ¤ Department of Physiology and Biophysics, School of Life Sciences, Fudan University, Shanghai 200433, China Received 21 January 2005, and in revised form 1 March 2005 Available online 23 March 2005
Abstract Human Xotillin-1 (reggie-2), a major hydrophobic protein of biomembrane microdomain lipid rafts, was cloned and expressed in Escherichia coli with four diVerent fusion tags (hexahistidine, glutathione S-transferase, NusA, and thioredoxin) to increase the yield. The best expressed Xotillin-1 with thioredoxin tag was solubilized from inclusion bodies, Wrst puriWed by immobilized metal aYnity column under denaturing condition and direct refolded on column by decreasing urea gradient method. The thioredoxin tag was cleaved by thrombin, and the Xotillin-1 protein was further puriWed by anion exchanger and gel Wltration column. The puriWed protein was veriWed by denaturing gel electrophoresis and Western blot. The typical yield was 3.4 mg with purity above 98% from 1 L culture medium. Using pull-down assay, the interaction of both the recombinant Xotillin-1 and the native Xotillin-1 from human erythrocyte membranes with c-Cbl-associated protein or neuroglobin was conWrmed, which demonstrated that the recombinant proteins were functional active. This is the Wrst report describing expression, puriWcation, and characterization of active recombinant raft speciWc protein in large quantity and highly purity, which would facilitate further research such as X-ray crystallography. 2005 Elsevier Inc. All rights reserved. Keywords: Flotillin-1 (reggie-2); Lipid rafts; Thioredoxin tag; c-Cbl-associated protein; Neuroglobin
Lipid rafts are newly discovered membrane microdomains enriched in sphingolipids and cholesterol. They are also called detergent-resistant membranes for their insolubility in most non-ionic detergents (reviewed in [1]). Many proteins involving in signal transduction and vesicular traYcking are abundant in lipid rafts, such as Src-family tyrosine kinases, protein kinase C, heterotrimeric and small G proteins, tyrosine kinase receptors, and G-protein-coupled receptors (reviewed in [2,3]). Besides these signaling proteins, lipid rafts also have some speciWc proteins, which were generally used as markers of lipid rafts, including Xotillin [4,5], caveolin [6], and stomatin [7]. But the structure, function, interac-
*
Corresponding author. Fax: +86 21 65650149. E-mail address:
[email protected] (Z. Zhang).
1046-5928/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.pep.2005.03.001
tion with other proteins, and linkage with disease of these proteins were not clearly deWned. Flotillins consist of two members: Xotillin-1 (reggie-2) and Xotillin-2 (reggie-1). They were originally discovered by their signiWcant upregulated expression in axonal regeneration after a lesion of the goldWsh optic nerve [4]. Bickel et al. [5] isolated a ‘new’ protein from the Triton X-100 insoluble membrane fraction of murine lung tissue in screening novel markers of lipid rafts, and named it “Xotillin” by their membrane state—Xoat like a Xotilla of ships in the Triton insoluble buoyant fraction. Sequence analysis shows that Xotillin-1 is identical to reggie-2 and Xotillin-2 is identical to reggie-1. Flotillin-1 and Xotillin-2 are homologous and evolutionary conserved [8]. Similar to stomatin, prohibitin, and bacterial membrane proteins HXK and HXC, Xottillins consist of conserved SPFH domain [9].
138
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
Human Xotillin-1 gene is single copy and has 13 exons, locates at 6p21.3. It is highly expressed in brain, heart, and lung. The human Xotillin-1 protein consists of 427 amino acid residues with predicted molecular weight of 47,355 Da, and predicted isoelectric point of 7.08 [10]. Flotillin-1 protein contains several potentially phosphorylated sites: residues 116 and 150 can be phosphorylated by casein kinase II, residues 160 and 238 by tyrosine kinase, and the residues 150 and 160 are evolutionary conserved. Flotillin-1 is also palmitoylated, which enables Xotillin-1 associate with plasma membrane through a conserved cysteine residue, Cys-34 [11]. On cell membrane, Xotillin-1, Xotillin-2, and caveolins exist as hetero-oligomeric complex [12]. Flotillin-1, Xotillin-2 and stomatin are the most abundant membrane proteins in erythrocyte lipid rafts. They present as independently organized high-order oligomers and act as separate scaVolding components at the cytoplasmic face of erythrocyte lipid rafts [7]. During malaria parasite Plasmodium falciparum infection, the association of Xotillins with erythrocyte lipid rafts was disrupted, and Xotillins were selectively recruited to the vacuole [13,14]. Flotillin-1 is associated with amyloid- protein, a protein resided in lipid rafts of human brain [15]. By Western blot and immunohistochemical analysis, Kokubo et al. [16] found high Xotillins expression in cortex during the development of senile plaque formation. Girardot et al. [17] also found that Xotillin-1 accumulated in neuron lysosomes in Alzheimer’s disease. Flotillin-1 also has important function in the second signaling pathway required for insulin-stimulated glucose transport. Apart from the phosphatidylinositol-3kinase-dependent pathway, c-Cbl-associated protein (CAP)1/Cbl complex can also Wrst form a ternary complex with Xotillin-1, then be localized to the plasma membrane lipid raft subdomain, Wnally stimulates glucose transport into fat and muscle cells [18]. Our research of searching the membrane proteins related to the type 2 diabetes in human erythrocyte by proteomics analysis also shows that Xotillin-1 is up-regulated in type 2 diabetes patients [19]. To do further study on the reason of abnormal Xotillin-1 expression in type 2 diabetes and the role of Xotillin-1 in glucose uptake, we want to get large quantity of functional Xotillin-1 protein. The lipid raft proteins were normally separated from membrane insoluble fraction in 1% Triton X-100 at low 1 Abbreviations used: BCA, bicinchoninic acid; BME, -mercaptoethanol; CAP, c-Cbl-associated protein; Chaps, 3-[(3-cholamidopropyl)dimethylamino]-1-propanesulfonate; DTT, dithiothreitol; EDTA, ethylenediaminetetraacetic acid; Flot1, Xotillin-1; GST, glutathione Stransferase; His6, hexahistidine; HRP, horseradish peroxidase; IPTG, isopropyl -D-thiogalactopyranoside; Ngb, neuroglobin; Ni–NTA, nickel–nitrilotriacetic acid; OG, n-octyl--D-glucoside; PMSF, phenylmethylsulfonyl Xuoride; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; SoHo, sorbin homology domain; TRX, thioredoxin.
temperature, with a further puriWcation step using sucrose gradient ultracentrifuge [5]. But it is hard to isolate Xotillin-1 from other lipid raft proteins. Bauer et al. [20] have isolated Xotillin-1 directly from goldWsh brain by two consecutive HPLC steps, the typical yield is 1 g Xotillin-1 from 30 goldWsh brains. The membrane association and hydrophobic character make Xotillin-1 protein puriWcation from tissues unfeasible. Researchers tried to get Xotillin-1 by express it in prokaryote [4,21], but there was no work involving puriWcation and characterization functional full-length Xotillin-1 in large quantity and high purity, so we tried to increase the expression level of Xotillin-1 by using diVerent tags, then to purify and characterize it. The biochemical activity of recombinant Xotillin-1 was compared with native Xotillin-1 extracted from human erythrocyte ghosts by pull-down assay. They all pulled-down by c-Cbl-associated protein or neuroglobin, both were known to interact with Xotillin1. The result showed that the recombinant and native Xotillin-1 were identical.
Materials and methods Materials The bacterial (Escherichia coli) hosts DH5, BL21 (DE3), HMS174 (DE3), the vectors pET21a, pET32a, pET43.1a, protein markers, and Western blot kits were obtained from Novagen (Madison, WI). KOD plus Pfu polymerase was purchased from Toyobo (Osaka, Japan). Nucleotides, agarose gel, DNA extraction kit, and high pure PCR puriWcation kit were purchased from Roche Diagnostics (Indianapolis, IN). Primers were synthesized at Bioasia (Shanghai, China). DNA sequencing was performed by Bioasia (Shanghai, China). The restriction endonucleases and DNA ligation kit were purchased from Takara (Dalian, China). Nickel–nitrilotriacetic acid (Ni–NTA) SuperXow column matrix was obtained from Qiagen (Chatsworth, CA). Sephacryl S-200 matrix, Sephadex G-50 matrix, Glutathione–Sepharose 4B, Thrombin, and pGEX-4T-1 vector were purchased from Amersham Biosciences (Piscataway, NJ). High Q Cartridge was from Bio-Rad (Hercules, CA). B-PER bacterial protein extraction reagent, bicinchoninic acid (BCA) protein assay reagent kit, goat anti-mouse secondary antibody labeled with horseradish peroxidase (HRP), and SuperSignal WestPico chemiluminescent substrate were from Pierce (Rockford, IL). MagneGST glutathione particles were from Promega (Madison, WI). Anti-Xotillin-1 monoclonal antibody was from BD Biosciences Pharmingen (San Diego, CA). n-Octyl--D-glucoside (OG) was from Dojindo Laboratories (Kumamoto, Japan). -Mercaptoethanol (BME), 3-[(3-cholamidopropyl)dimethylamino]-1-propanesulfonate (Chaps), dithiothreitol (DTT), ethylenediaminetetraacetic acid (EDTA),
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
imidazole, isopropyl -D-thiogalactopyranoside (IPTG), L-glutathione (reduced form), phenylmethylsulfonyl Xuoride (PMSF), Triton X-100, Triton X-114, Tween 20, SB3-10, SB3-14, and SB3-16 were from Sigma (St. Louis, MO). Amicon Ultra-15 centrifugal Wlter (MWCO 10,000) was obtained from Millipore (Bedford, MA). Human bloods were obtained from the Shanghai Red Cross Blood Center. All other reagents were of analytical grade.
139
fresh LB medium and grew in 37 °C shaker (240 rpm) to OD (600 nm) of 0.6, then the cells were cooled down to 25 °C. Expression of the Xotillin-1 fusion proteins was induced for 5 h by adding IPTG to a Wnal concentration of 0.1 mM. The cells were collected by centrifugation at 6000g for 15 min and were frozen at ¡20 °C until use. PuriWcation and refold of Xotillin-1 fusion protein with thioredoxin tag
Cloning the Xotillin-1 coding region into vector pT7470 The pT7470 expression vector containing both N-terminus and C-terminus hexahistidine (His6) tag was constructed from vector pET21a, the detailed pT7470 vector construction procedure was described previously [22]. Human embryo brain cDNA was used as nested PCR’s template, which was kindly provided by Dr. Ping Xu, Laboratory of Genomic Physiology, School of Life Sciences, Fudan University. The primers used were listed in Table 1. The product of second round PCR was double digested by EcoRI and XhoI, ligated with the double digested expression vector pT7470 and transformed into E. coli host DH5. The positive pT7470-Xotillin-1 (Flot1) clone was selected and sequenced for veriWcation. Cloning the Xotillin-1 coding region into other expression vectors pT7470-Flot1 plasmid was used to clone Xotillin-1 into other vectors. The pT7470-Flot1 plasmid was double digested by EcoRI and XhoI, the Xotillin-1 fragment was ligated with EcoRI and XhoI double digested pET32a, pGEX-4T-1, and pET43.1a vectors, then transformed into E. coli host DH5, and the positive clone was selected and sequenced. Expression four types of Xotillin-1 fusion proteins The expression vectors of pT7470-Flot1, pET32aFlot1, pGEX-Flot1, and pET43.1a-Flot1 were transformed into BL21 (DE3) or HMS174 (DE3) for expression. After grown overnight at 37 °C in 20 ml LB medium supplemented with 100 g/ml ampicillin, a portion (2 ml) of the bacterial suspension was then transferred into 1 L Table 1 Sequences of synthesized oligonucleotide primers used in cloning the Xotillin-1 gene into vector pT7470 Primer
Sequencea
P1 P2 P3 P4
5⬘-CTCCCAGGAAAGCTGGTCTG-3⬘ 5⬘-GCATCTGTGAGGGCTGAAGG-3⬘ 5⬘-AAAGAATTCATGTTTTTCACTTGT-3⬘ 5⬘-ATACTCGAGTCAGGCTGTTCTCAA-3⬘
a P1 and P2 were used in the Wrst round of nested PCR, P3 and P4 were used in the second round of nested PCR. The sequences cut by respective restriction endonucleases (P3, EcoRI; P4, XhoI) are underlined.
The cells were thawed and resuspended in 200 ml buVer A (50 mM Tris–HCl, pH 8.0, 100 mM NaCl, and 200 g/ml PMSF) and were lysed by sonication. The cellular debris was collected by centrifugation at 20,000g for 20 min and washed by buVer B (50 mM Tris–HCl, pH 8.0, 1% Triton X-100, and 200 g/ml PMSF) for 3 times. Finally, the washed inclusion bodies were solubilized in 40 ml buVer C (8 M Urea, 50 mM Tris–HCl, pH 8.0, 100 mM NaCl, 1 mM BME, and 200 g/ml PMSF) containing 10 mM imidazole with gentle circular agitation for 2 h at 4 °C. After centrifugation at 20,000g for 20 min, the supernatant fraction was collected and loaded on a 10 ml (15 £ 1 cm) Ni–NTA SuperXow column already equilibrated with buVer C containing 10 mM imidazole. Then the column was washed by 20 ml buVer C containing 20 mM imidazole and was ready for the refolding on column. The linear gradient formed with buVer C containing 10 mM imidazole (100–0%) and buVer D (50 mM Tris–HCl, pH 8.0, 100 mM NaCl, 0.1% OG, 1% Tween 20, 1 mM BME, and 200 g/ml PMSF) containing 10 mM imidazole (0–100%), the Xow rate was 0.5 ml/min and total volume was 100 ml. The refolded TRX-Flot1 protein was eluted with buVer D containing 300 mM imidazole and loaded on a Sephadex G-50 column equilibrated with buVer D (without PMSF) to remove imidazole and PMSF. Thioredoxin (TRX) tag was digested by thrombin (5 U/ml) at 8 °C for 12 h. The TRX tag and uncleaved TRX-Flot1 were removed by passing through Ni–NTA SuperXow column again. The sample was loaded on a 5 ml High Q anion exchange column equilibrated with buVer D and eluted by the linear gradient formed by buVer D and buVer D containing 300 mM NaCl in Xow rate of 0.5 ml/min and total volume of 20 ml. The Xotillin1 peak was collected and concentrated using Amicon Ultra-15 centrifugal Wlter before loading on Sephacryl 200 column (60 £ 1 cm) equilibrated with buVer E (50 mM Tris–HCl, pH 8.0, 100 mM NaCl, 1% OG, 1 mM DTT, 1 mM EDTA, and 200 g/ml PMSF). The peak was collected, concentrated, and stored at 4 °C. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and Western blot Ten microliter protein samples were mixed with 5 l 3£ loading buVer [150 mM Tris–HCl (pH 6.8), 300 mM DTT, 6%(w/v) SDS, 0.06%(w/v) bromophenol blue, and
140
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
30%(v/v) glycerol], and boiled for 3 min prior to load on a 10 or 12% Tris–HCl gel for electrophoresis. Proteins in the gel were visualized by staining with Coomassie G-250. The purity of stained proteins was determined by BioRad’s Quantity One 1-D gel analysis software. The protein concentration was determined by BCA method according to the protocol of BCA Protein Assay Reagent Kit (Pierce). For Western blot analysis, the proteins were electrophoretically transferred onto 0.20 m nitrocellulose transfer membrane (Schleicher & Schuell). The nitrocellulose was blocked by incubation in blocking buVer (150 mM NaCl, 50 mM Tris–HCl, pH 7.4, and 5%(w/v) low fat milk powder) for 1 h at 37 °C, then incubated with anti-Xotillin-1 monoclonal antibody (1:2000) in the blocking buVer for 1 h at 37 °C followed by washing four times with washing buVer (150 mM NaCl, 50 mM Tris–HCl, pH 7.4, and 0.1%(v/v) Tween 20). The membrane was then incubated with goat anti-mouse secondary antibody conjugated with HRP (1:5000) for 1 h at 37 °C. Also, the unbound HRP-conjugated secondary antibody was removed by washing with washing buVer four times for 10 min each. Finally, the membrane was incubated with working solution (SuperSignal WestPico Substrate) for 1 min and then pressed against an autoradiographic Wlm. Cloning, expressing, and purifying glutathione S-transferase-CAP sorbin homology domain (SoHo) and GST-neuroglobin (Ngb) Human embryo brain cDNA was used as nested PCR’s template to clone the human CAPSoHo domain and neuroglobin. The primers used were listed in Table 2. The products of the second round PCR reaction were puriWed, double digested, ligated to the vector pGEX4T-1, and transformed into E. coli host DH5. The positive clone was selected and sequenced. The vectors pGEX-CAPSoHo and pGEX-Ngb were then transformed into BL21 (DE3), the proteins were expressed at 25 °C with 0.1 mM IPTG induction for 5 h, and puriWed Table 2 Sequences of synthesized oligonucleotide primers used in cloning the CAPSoHo (P1–P4) and neuroglobin (P5–P8) gene into vector pGEX4T-1 Primer
Sequencea
P1 P2 P3 P4 P5 P6 P7 P8
5⬘-GCAGACGACTTGTCCTGCCACC-3⬘ 5⬘-TGAATGATGCTTCCTCCTCCAA-3⬘ 5⬘-GAAGAATTCGTGGGAGAGCAGGAC-3⬘ 5⬘-ATGCTCGAGTTATGGGAGGGGTAGTGT-3⬘ 5⬘-GGGTGTCTCCACCTACGA-3⬘ 5⬘-CCACCTCTGCCACCAAA-3⬘ 5⬘-CGGGGATCCATGGAGCGCCCGGAG-3⬘ 5⬘-CGGGGTCTCGAGTTACTCGCCATCCCA-3⬘
P1, P2 and P5, P6 were used in the Wrst round of nested PCR, P3, P4 and P7, P8 were used in the second round of nested PCR. The sequences cut by respective restriction endonucleases (P3, EcoRI; P4, XhoI; P7, BamHI; and P8, XhoI) are underlined. a
as the AP biotech’s pGEX system protocol. The proteins eluted from Glutathione–Sepharose 4B column were further puriWed by High Q anion exchanger. The vector pGEX-4T-1 was transformed into BL21 (DE3), the control glutathione S-transferase (GST) protein was expressed and puriWed. All the puriWed GST fusion proteins were concentrated by Amicon Ultra-15 centrifugal Wlter and stored at 4 °C. Extraction Xotillin-1 from the human erythrocytes Extraction of Xotillin-1 from the erythrocytes was performed according to an established method [7] with minor modiWcations. BrieXy, the fresh human erythrocytes were washed with PBS (155 mM NaCl and 5 mM sodium phosphate, pH 7.4) and hemolyzed in 20£ (vol) 5 mM sodium phosphate buVer, pH 7.4. After washing several times with 5 mM sodium phosphate buVer, the erythrocyte ghosts (the pellets) were treated with 1% Triton X-100 in PBS with 200 g/ml PMSF at 4 °C for 30 min. The pellets were resuspended in cold 60% sucrose and 1% Triton X-100 in PBS, with a Wnal sucrose concentration of 40%, placed in centrifuge tubes, overlayed with a linear 10–30% sucrose in PBS, and centrifuged in a SW41 rotor (Beckman) at 200,000g for 16 h. Lipid raft fractions were pooled, treated with buVer E (50 mM Tris–HCl, pH 8.0, 100 mM NaCl, 1% OG, 1 mM DTT, 1 mM EDTA, and 200 g/ml PMSF) at 4 °C for 30 min, centrifuged at 20,000g for 10 min. The soluble fraction was collected and used for pull-down assay. Pull-down assay The puriWed GST-CAPSoHo (15 g), GST-Ngb protein (15 g) or control GST protein (10 g) were loaded on the MagneGST glutathione particles (50 l of 50% slurry) equilibrated with buVer E and blocked by bovine serum albumin (100 g). One hundred micrograms of the target protein (thrombin digested recombinant Xotillin-1 or puriWed raft protein including native Xotillin-1 extracted from the erythrocytes) was loaded on the resin. The mixture was gently shaken overnight at 4 °C, and washed four times with buVer E, 10 min each. Proteins were eluted by SDS–PAGE loading buVer and analyzed by SDS–PAGE followed by Western blot analysis.
Results and discussion Cloning the Xotillin-1 coding region into vector pT7470 Full-length Xotillin-1 coding sequence from reverse transcripted cDNA was directly cloned to pT7470 containing His6 tag. Since ineYciency of ordinary PCR, two sets of primers were synthesized for nested PCR. The sequence was veriWed. Then His6-Flot1 was expressed in
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
141
Fig. 1. SDS–PAGE and Western blot analysis of the expression of recombinant Xotillin-1 fusion proteins with diVerent tags. (A) SDS–PAGE patters of the Xotllin-1 with four diVerent tags. (B) Western blot result of the four recombinant proteins using anti-Xotillin-1 antibody after 15 s exposure. (C) Western blot result of the four recombinant proteins after 10 min exposure. Marker, Novagen’s Perfect Protein markers; (¡), uninduced bacterial lysate; (+)S, soluable fraction of B-PER reagent extraction of induced bacterial lysate; (+)I, insoluable fraction of B-PER extraction of induced bacterial lysate. S-Protein HRP conjugate was used for detecting the Novagen’s Perfect Protein markers in Western blot analysis.
BL21 (DE3) by IPTG induction. The yield is low and most expressed His6-Flot1 was in inclusion body (Fig. 1). The factors of time, temperature, IPTG concentration, and diVerent hosts of HMS174 (DE3) were tested and there was no signiWcant improvement. Then diVerent vectors were tested to get better expression. Expression and comparison of four Xotillin-1 fusion proteins Four most frequently used prokaryotic expression tags were used to improve the expression and solubility
of Xotillin-1. Except of the His6 tag included in pT7470 vector mentioned above, fusion tags of TRX, GST, and NusA were tried. TRX was relatively small (<12 kDa) as compare to GST and NusA. The crystal structure of TRX showed that the N- and C-termini of TRX are accessible on the molecule’s surface, and when TRX was overexpressed in E. coli, it can accumulate to 40% of the total cellular proteins and remained in the soluble fraction [23]. We Wrst tried to use the pET32a vector that included TRX tag to express Xotillin-1. The expression of Xotillin-1 with Schistosoma japonicum GST (around 26 kDa) tag [24] in pGEX-4T-1 vector or with bigger
142
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
NusA (around 55 kDa) [25] tag in pET43.1a vector is also checked. These tags can be conveniently cleaved by speciWc proteases (such as thrombin, enterokinase or Factor Xa). The induced cells were lysed by B-PER reagent and separated as soluble and insoluble fraction. The expression levels of the four fusion proteins were visualized by SDS–PAGE and Western blot as shown in Fig. 1. Because the expression levels of Xotillin-1 fusion proteins varied markedly, exposure one time can only get partial information of the expression level. One Wlm was exposed 15 s (Fig. 1B) to see the best expressed proteins and then over-exposed 10 min (Fig. 1C) to get the detailed information. Compared with the theoretical molecular weight of the four proteins listed in Table 3, SDS–PAGE (Fig. 1A) showed that the TRX-Flot1 or GST-Flot1 expression was better than His6-Flot1 expression. The NusA-Flot1 vector was expressed at around 75 kDa. As compared with the theoretical predicted 108 kDa, the expressed protein looks like somehow in truncated form. Western blot patterns of the four proteins by anti-Xotillin-1 antibody were compared in Figs. 1B and C. His6-Flot1 was detected in the insoluble fraction, but the expression level was low. TRX-Flot1 was detected in both supernatant and insoluble fractions and there were more target proteins in insoluble fraction. GST-Flot1 was detected only in insoluble fraction. In Western bolt analysis, there were only few full-length NusAFlot1 and most of them were truncated, showing a band about 45 kDa in Fig. 1B. After long exposure (Fig. 1C), there were several faint bands between 45 kDa and full-length 108 kDa. We tried to increase the soluble fraction of TRXFlot1 and GST-Flot1. After decreased the IPTG concentration to 0.1 mM and lower the temperature to 20 °C, the soluble fraction did not increase dramatically (less than 10% of total target protein, data not shown). In further research, urea was used to solubilize the target protein in inclusion body. It was found that the glutathione beads cannot bind inactive form of GST fusion protein in high concentration denaturant, but the Ni–NTA beads can bind denatured TRXFlot1, so TRX-Flot1 was chosen for further puriWcation. Table 3 The number of amino acid residues, the theoretical molecular weight (Mw) and isoelectric point (pI) of the native human Xotillin-1 and four recombinant Xotillin-1 fusion proteins Native Flot1 His6-Flot1 TRX-Flot1 GST-Flot1 NusA-Flot1
Number of amino acid
Mw (Da)
pI
427 443 594 657 982
47355.2 49282.3 65334.5 74150.4 108389.7
7.08 6.92 5.96 6.31 5.06
Optimizing the puriWcation and refolding procedure of TRX-Flot1 fusion protein TRX-Flot1 fusion protein was induced in diVerent temperatures (from 37 to 16 °C) and diVerent IPTG concentrations (from 2 to 0.05 mM). The expression was signiWcant but most of them were always in the inclusion body. Typically there were around 140 mg TRX-Flot1 expressed when induced with 0.1 mM IPTG at 25 °C for 5 h. But only about 0.2 mg soluble TRX-Flot1 can be obtained from 1 L culture medium by a series of chromatography (data not shown). To get enough recombinant Xotillin-1 for crystallography study, the recombinant proteins in the inclusion body were extracted with a strong denaturant urea followed by three chromatographic steps: an aYnity column (Ni–NTA superXow) for refolding and primary puriWcation; an anion exchange column (High Q) for further puriWcation; and a size exclusion chromatography (Sephacryl S-200) step to maximize the purity. When the denaturant was removed by traditional dialysis method, TRX-Flot1 lower than 0.05 mg/ml still precipitated. So refolding TRX-Flot1 using Sephacryl S200 gel Wltration chromatography was tried [26]. Though refolding by gel Wltration minimized the dilution of TRX-Flot1, it still precipitated because the removal of denaturant was fast. TRX-Flot1 also had His6 tag and bound Ni–NTA under both denatured form and native form, so Ni–NTA was used to immobilize the TRXFlot1 and then TRX-Flot1 was refolded by gradually decreasing the concentration of denaturant. BME was used in Ni–NTA puriWcation instead of DTT and EDTA was added immediately in eluate. We have tried to cleave the TRX tag by thrombin at room temperature. More than 95% of TRX-Flot1 can be cleaved in 2 h at 25 °C. But most of the cleaved recombinant Xotillin-1 proteins were aggregated and precipitated. Perhaps the short time of high temperature cleavage was not enough for hydrophobic Xotillin1 correct refolding when the soluble thioredoxin tag was removed. After a series of time and temperature tested, we Wnally cleaved TRX-Xotillin-1 by thrombin at 8 °C for 12 h. We also tried to do on column thrombin cleavage while the refolded TRX-Flot1 was still on Ni– NTA superXow column. If it works, buVer without imidazole can elute the cleaved Xotillin-1 (without His6 tag) while cleaved thioredoxin tag and uncut TRXFlot1 (both contain His6 tag) still bind to the resin. But for unknown reason, the thrombin cleavage on column was ineYcient. We also compared High Q anion exchanger with phenyl-Sepharose hydrophobic interaction chromatography in further puriWcation of Xotillin-1, and the resolution of High Q anion exchanger was much better than phenyl-Sepharose hydrophobic interaction. Since the interaction of Xotillin-1 with High Q was not strong,
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
Xotillin-1 was eluted at moderate salt concentration (0.2– 0.3 M NaCl), so the salt should be removed before loaded on High Q column, and we passed the protein solution through a Sephadex G-50 column to desalt. We did not choose cation exchanger because this method would decrease the buVer’s pH to lower than 5.0, far away from the physiological pH. When there was not enough reductant added in the puriWcation, lots of the cleaved recombinant Xotillin-1 and fusion protein TRX-Flot1 existed as trimers. This can be proved by the Xotillin-1 gel Wltration elution position and mobility shift in non-reducing SDS–PAGE (data not shown). Sometimes even with low concentration reducing agents, the disulWde bridge cannot be eYciently broken, and there still existed a few Xotillin-1 trimer portions.
143
The solubility of human erythrocyte Xotillin-1 was tested in most used non-ionic detergents and their combination, including OG, Triton X-100, Triton X-114, Tween 20, Chaps, SB3-10, SB3-14, and SB3-16. It was found that Xotillin-1 was most soluble in OG, especially when OG’s concentration reached 1% (data not shown). We Wnally used low concentration of OG (0.1%) at the beginning of puriWcation and used 1% OG at the gel Wltration step for preparing large quantity of Xotillin-1. Table 4 and Fig. 2 summarized the whole puriWcation procedure of recombinant Xotillin-1. The typical yield of recombinant Xotillin-1 from TRX-Flot1 was about 3.4 mg from 1 L culture medium, the purity was more than 98%, that was enough to do most traditional biochemical experiments.
Table 4 PuriWcation of recombinant Xotillin-1 protein from an E. coli expression systema Inclusion body Ni–NTA High Q Sephacryl S-200
Total protein (mg)
Flotillin-1 (mg)
Fold enrichment
Recovery (%)
192.8 15.0 7.03 3.44
110.9 (140.9b) 10.7 (13.6b) 6.54 3.41
1 1.24 1.62 1.72
100 9.6 5.9 3.1
a
From 1 L of E. coli culture medium. The value was calculated according to the molecular weight before (TRX-Flot1, 65.3 kDa) and after thrombin cleavage (Xotillin-1, 51.4 kDa), the value in parentheses was the origin data gained from gel band density of 65.3 kDa TRX-Flot1 protein. b
Fig. 2. SDS–PAGE and Western blot analysis of the puriWcation procedure of recombinant Xotillin-1. Lane 1, Novagen’s Perfect Protein markers; lane 2, induced whole bacterial lysate; lane 3, the proteins loaded on Ni–NTA column (washed and solubilized fusion protein TRX-Flot1 from inclusion body); lane 4, the primary puriWed and refolded TRX-Flot1 eluted from Ni–NTA aYnity chromatography; lane 5, the eluate from High Q anion exchanger after the thioredoxin tag was digested by thrombin; lane 6, the puriWed recombinant Xotillin-1 by Sephacryl S-200 gel Wltration chromatography; and lanes 7–12 represent the Western blot results of lanes 1–6 using anti-Xotillin-1 antibody after 15 s exposure.
144
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145
Biochemical characterization Xotillin-1 by pull-down assay with GST-CAPSoHo and GST-Ngb There is no report to show that Xotillin-1 has some enzyme activity while Xotillin-1 was reported to interact with several other proteins, so we decided to compare the native and recombinant Xotillin-1 by pull-down assay using known Xotillin-1 associated proteins. C-Cbl-associated protein (Swiss-Prot: Q9BX66) is also called “Sorbin and SH3 domain-containing protein” because it contains a gut peptide sorbin homology region in N-terminal and three SH3 domains in C-terminal [27]. CAP was a multifunctional adaptor protein that interacts with many other proteins. During insulin stimulation, CAP dissociates from the insulin receptor, then Xotillin-1 interacts with CAP somehow and the CAP/ Cbl complex migrates to lipid rafts [18]. By CAP gene deletion analyses, Kimura et al. [28] identiWed a 115-aa sorbin homology domain binding with Xotillin-1. Neuroglobin (Swiss-Prot: Q9NPG2) was discovered in vertebrates and predominantly expressed in the brain and other nerve tissues, it shares little sequence similarity with hemoglobin and myoglobin [29]. Wakasugi et al. [21] found that Xotillin-1 interacted with neuroglobin by screening a human brain cDNA library in yeast twohybrid system, and the interaction was conWrmed by pull-down assay with truncated Xotillin-1. So, the interaction with CAP or neuroglobin was chosen to test the functional activity of recombinant Xotillin-1 in this study by pull-down assay. Expression of full-length CAP protein with GST tag was tried, but most of the induced proteins were the truncated form. So the sorbin homology domain of CAP with a N-terminal GST tag was expressed, and enough recombinant GST-CAPSoHo protein was obtained. To further purify of GST-CAPSoHo and GST-Ngb from truncated GST protein, one more column (High Q anion exchanger) was used. Fig. 3 showed the association of Xotillin-1 with GSTCAPSoHo or GST-Ngb by pull-down assay. Both the recombinant Xotillin-1 and the native Xotillin-1 bound to GST-CAPSoHo or GST-Ngb, but not to control GST protein. It conWrms that the puriWed recombinant Xotillin-1 was biochemical similar with the native Xotillin-1. As we have mentioned that Xotillin-1 protein is palmitoylated in vivo and contains several potentially phosphorylated sites, it will be interesting to address in the future whether these or other posttranslational modiWcations in mammalian cells can inXuence the function of recombinant Xotillin-1. In conclusion, we have cloned, expressed, puriWed, and characterized a soluble and functional human Xotillin-1, a hydrophobic lipid raft protein. The soluble recombinant Xotillin-1 is now being used in aYnity chromatography for the isolation of Xotillin-1 binding proteins, in in vitro studies to investigate raft protein
Fig. 3. Association of Xotillin-1 with GST-CAPSoHo or GST-Ngb by pull-down assay. The puriWed GST negative control (lanes 3 and 7, 10 g each), GST-CAPSoHo (lanes 4 and 8, 15 g each) or GST-Ngb (lanes 5 and 9, 15 g each) was loaded on the MagneGST glutathione particles balanced with buVer E (50 mM Tris–HCl, pH 8.0, 100 mM NaCl, 1% OG, 1 mM DTT, 1 mM EDTA, and 200 g/ml PMSF). The target protein, puriWed recombinant Xotillin-1 (lanes 3–5, 100 g each) or puriWed raft protein including native Xotillin-1 extracted from the erythrocytes (lanes 7–9, 100 g raft protein each), were loaded on the resin with gently shaking over night at 4 °C. After washing with buVer E, the proteins eluted by SDS–PAGE loading buVer were analyzed by SDS–PAGE followed with Western blot using anti-Xotillin-1 antibody. Lane 1, Novagen’s Perfect Protein markers; lane 2, positive control (+) of puriWed recombinant Xotillin-1; and lane 6, positive control (+) of raft protein including native Xotillin-1 extracted from erythrocytes.
interaction, in a cell-free screening system for the identiWcation of type 2 diabetes related proteins, and also in X-ray crystallography analysis.
Acknowledgments The work was supported by the Grant 30370380 from the National Natural Science Foundation of China and the Grant 2001CB5102 from the National Basic Research Priorities Program of China.
References [1] D.A. Brown, E. London, Functions of lipid rafts in biological membranes, Annu. Rev. Cell Dev. Biol. 14 (1998) 111–136. [2] R.G.W. Anderson, The caveolae membrane system, Annu. Rev. Biochem. 67 (1998) 199–225. [3] M. Edidin, The state of lipid rafts: from model membranes to cells, Annu. Rev. Biophys. Biomol. Struct. 32 (2003) 257–283. [4] T. Schulte, K.A. Paschke, U. Laessing, F. Lottspeich, C.A. Stuermer, Reggie-1 and reggie-2, two cell surface proteins expressed by retinal ganglion cells during axon regeneration, Development 124 (1997) 577–587. [5] P.E. Bickel, P.E. Scherer, J.E. Schnitzer, P. Oh, M.P. Lisanti, H.F. Lodish, Flotillin and epidermal surface antigen deWne a new family of caveolae-associated integral membrane proteins, J. Biol. Chem. 272 (1997) 13793–13802. [6] P.S. Liu, M. Rudick, R.G.W. Anderson, Multiple functions of caveolin-1, J. Biol. Chem. 277 (2002) 41295–41298.
Y. Ding et al. / Protein Expression and PuriWcation 42 (2005) 137–145 [7] U. Salzer, R. Prohaska, Stomatin, Xotillin-1, and Xotillin-2 are major integral proteins of erythrocyte lipid rafts, Blood 97 (2001) 1141–1143. [8] E. Malaga-Trillo, U. Laessing, D.M. Lang, A. Meyer, C.A. Stuermer, Evolution of duplicated reggie genes in ZebraWsh and GoldWsh, J. Mol. Evol. 54 (2002) 235–245. [9] N. Tavernarakis, M. Driscoll, N.C. Kyrpides, The SPFH domain: implicated in regulating targeted protein turnover in stomatins and other membrane-associated proteins, Trends Biochem. Sci. 24 (1999) 425–427. [10] A.J. Edgar, J.M. Polak, Flotillin-1: gene structure: cDNA cloning from human lung and the identiWcation of alternative polyadenylation signals, Int. J. Biochem. Cell Biol. 33 (2001) 53–64. [11] I.C. Morrow, S. Rea, S. Martin, I.A. Prior, R. Prohaska, J.F. Hancock, D.E. James, R.G. Parton, Flotillin-1/reggie-2 traYcs to surface raft domains via a novel Golgi-independent pathway, J. Biol. Chem. 277 (2002) 48834–48841. [12] D. Volonte, F. Galbiati, S. Li, K. Nishiyama, T. Okamoto, M.P. Lisanti, Flotillins/cavatellins are diVerentially expressed in cells and tissues and form a hetero-oligomeric complex with caveolins in vivo, J. Biol. Chem. 274 (1999) 12702–12709. [13] E. Nagao, K.B. Seydel, J.A. Dvorak, Detergent-resistant erythrocyte membrane rafts are modiWed by a Plasmodium falciparum infection, Exp. Parasitol. 102 (2002) 57–59. [14] S.C. Murphy, B.U. Samuel, T. Harrison, K.D. Speicher, D.W. Speicher, M.E. Reid, R. Prohaska, P.S. Low, M.J. Tanner, N. Mohandas, K. Haldar, Erythrocyte detergent-resistant membrane proteins: their characterization and selective uptake during malarial infection, Blood 103 (2004) 1920–1928. [15] S.J. Lee, U. Liyanage, P.E. Bickel, W. Xia, P.T. Lansbury Jr., K.S. Kosik, A detergent-insoluble membrane compartment contains a beta in vivo, Nat. Med. 4 (1998) 730–734. [16] H. Kokubo, C.A. Lemere, H. Yamaguchi, Localization of Xotillins in human brain and their accumulation with the progression of Alzheimer’s disease pathology, Neurosci. Lett. 290 (2000) 93–96. [17] N. Girardot, B. Allinquant, C. Duyckaerts, Lipid rafts, Xotillin-1 and Alzheimer disease, J. Soc. Biol. 197 (2003) 223–229. [18] C.A. Baumann, V. Ribon, M. Kanzaki, D.C. Thurmond, S. Mora, S. Shigematsu, P.E. Bickel, J.E. Pessin, A.R. Saltiel, CAP deWnes a second signalling pathway required for insulin-stimulated glucose transport, Nature 407 (2000) 202–207.
145
[19] M. Jiang, L. Jia, W. Jiang, X. Hu, H. Zhou, X. Gao, Z. Lu, Z. Zhang, Protein disregulation in red blood cell membranes of type 2 diabetic patients, Biochem. Biophys. Res. Commun. 309 (2003) 196–200. [20] S.H. Bauer, M.F. Wiechers, K. Bruns, M. Przybylski, C.A. Stuermer, Isolation and identiWcation of the plasma membrane-associated intracellular protein reggie-2 from goldWsh brain by chromatography and Fourier-transform ion cyclotron resonance mass spectrometry, Anal. Biochem. 298 (2001) 25–31. [21] K. Wakasugi, T. Nakano, C. Kitatsuji, I. Morishima, Human neuroglobin interacts with Xotillin-1, a lipid raft microdomain-associated protein, Biochem. Biophys. Res. Commun. 318 (2004) 453–460. [22] Y. Ding, W.H. Jiang, Y. Su, H.Q. Zhou, Z.H. Zhang, Expression and puriWcation of recombinant cytoplasmic domain of human erythrocyte band 3 with hexahistidine tag or chitin-binding tag in Escherichia coli, Protein Expr. Purif. 34 (2004) 167–175. [23] E.R. Lavallie, E.A. Diblasio, S. Kovacic, K.L. Grant, P.F. Schendel, J.M. Mccoy, A thioredoxin gene fusion expression system that circumvents inclusion body formation in the Escherichia coli cytoplasm, Biotechnology 11 (1993) 187–193. [24] W.G. Kaelin, W. Krek, W.R. Sellers, J.A. Decaprio, F. Ajchenbaum, C.S. Fuchs, T. Chittenden, Y. Li, P.J. Farnham, M.A. Blanar, D.M. Livingston, E.K. Flemington, Expression cloning of a cDNA encoding a retinoblastoma-binding protein with E2F-like properties, Cell 70 (1992) 351–364. [25] G.D. Davis, C. Elisee, D.M. Newham, R.G. Harrison, New fusion protein systems designed to give soluble expression in Escherichia coli, Biotechnol. Bioeng. 65 (1999) 382–388. [26] M.H. Werner, G.M. Clore, A.M. Gronenborn, A. Kondoh, R.J. Fisher, Refolding proteins by gel-Wltration chromatography, FEBS Lett. 345 (1994) 125–130. [27] W.H. Lin, C.J. Huang, M.W. Liu, H.M. Chang, Y.J. Chen, T.Y. Tai, L.M. Chuang, Cloning, mapping, and characterization of the human sorbin and SH3 domain containing 1 (SORBS1) gene: a protein associated with c-Abl during insulin signaling in the Hepatoma cell line Hep3B, Genomics 74 (2001) 12–20. [28] A. Kimura, C.A. Baumann, S.H. Chiang, A.R. Saltiel, The sorbin homology domain: a motif for the targeting of proteins to lipid rafts, Proc. Natl. Acad. Sci. USA 98 (2001) 9098–9103. [29] T. Burmester, B. Weich, S. Reinhardt, T. Hankein, A vertebrate globin expressed in the brain, Nature 407 (2000) 520–523.