Enhanced expression and purification of inositol 1,4,5-trisphosphate 3-kinase A through use of the pCold1-GST vector and a C-terminal hexahistidine tag in Escherichia coli

Protein Expression and Purification 97 (2014) 72–80

Contents lists available at ScienceDirect

Protein Expression and Purification journal homepage: www.elsevier.com/locate/yprep

Enhanced expression and purification of inositol 1,4,5-trisphosphate 3-kinase A through use of the pCold1-GST vector and a C-terminal hexahistidine tag in Escherichia coli Dongmin Lee, Seungrie Han, Seungkyun Woo, Hyun Woo Lee, Woong Sun, Hyun Kim ⇑ Department of Anatomy, College of Medicine, Korea University, Brain Korea 21, Seoul 136-705, Republic of Korea

a r t i c l e

i n f o

Article history: Received 10 October 2013 and in revised form 13 February 2014 Available online 25 February 2014 Keywords: Bacterial expression system Inositol 1,4,5-trisphosphate 3-kinase A Histidine-mediated affinity purification pCold vector Recombinant protein

a b s t r a c t Inositol 1,4,5-trisphosphate 3-kinase A (IP3K-A, alternative name: ITPKA) is a neuron-specific enzyme that converts 1,4,5-trisphosphate (IP3) into inositol 1,3,4,5-tetrakisphosphate (IP4) through its kinase domain. In addition, transient overexpression of IP3K-A induces morphological changes in dendritic spines of excitatory synapses in a kinase-independent manner, apparently by modulating the organization of the neuronal cytoskeleton. Although the procurement of a purified recombinant IP3K-A protein would be indispensable for the biochemical elucidation of its physiological roles, production of recombinant IP3K-A has proven technically challenging in conventional Escherichia coli expression systems. These difficulties stem from low enzyme solubility, as well as poor protein quality caused by the tendency of IP3K-A to split into partial fragments. In present study, we newly introduced cold-shock expression vector (pCold1) together with a C-terminal hexahistidine tag (C-HIS) to enhance the expression levels of recombinant IP3K-A in E. coli. Importantly, when compared with other commonly-employed bacterial expression systems, the pCold1 system improved the yield and the purity of full-length IP3K-A due to the exclusion of truncated enzyme forms, and also enhanced the solubility of the enzyme. Furthermore, the functional integrity of purified IP3K-A was confirmed in both kinase activity assay and microtubule binding assay. Recombinant IP3K-A acquired via this modified protocol will be expected to facilitate the exploration of the enzyme’s biochemical profile, both structurally and functionally. Ó 2014 Elsevier Inc. All rights reserved.

Introduction Inositol 1,4,5-trisphosphate 3-kinase A (IP3K-A)1 is a neuronspecific protein that is mainly expressed in dendrites and small dendritic protrusions, termed dendritic spines [1,2]. IP3K-A was first identified due to its kinase activity, which converts inositol 1,4,5-trisphosphate (IP3) into inositol 1,3,4,5-tetrakisphosphate (IP4). The neuronal IP3/IP4 balance is critical for the regulation of calcium homeostasis and signaling, because IP3 has the capacity to increase intracellular calcium levels by opening the endoplasmic ⇑ Corresponding author. Tel.: +82 2286 1153; fax: +82 929 5696. E-mail address: [email protected] (H. Kim). Abbreviations used: C-HIS, C-terminal hexahistidine tag; CBB, Coomassie brilliant blue; CV, column volume; DTT, dithiothreitol; FW, forward; GST, glutathione S-transferase; HRV, human rhinovirus; IP3, 1,4,5-trisphosphate; IP3K-A, inositol 1,4,5-trisphosphate 3-kinase A; IP4, inositol 1,3,4,5-tetrakisphosphate; IPTG, b-D1-thiogalactopyranoside; MBP, maltose binding protein; N-HIS, N-terminal hexahistidine tag; OD, optical density; PCR, polymerase chain reaction; RV, reverse; SDS, sodium dodecyl sulfate; SDS–PAGE, SDS–polyacrylamide gel electrophoresis; TEV, tobacco etch virus. 1

http://dx.doi.org/10.1016/j.pep.2014.02.006 1046-5928/Ó 2014 Elsevier Inc. All rights reserved.

reticulum gate. The net result is the release of calcium reservoirs into the cytoplasm. Therefore, IP3K-A contributes to calcium signaling by reducing the intraneuronal levels of IP3 in favor of IP4. Structurally, IP3K-A is composed of a N-terminal regulatory domain, a central calmodulin binding domain, and a C-terminal kinase domain [3]. The structure of the catalytic domain (amino acids 185–459) was previously determined by two different research groups [4,5]. However, the structure of full-length IP3K-A, including the N-terminal regulatory domain, is still unknown due to the overall instability and flexibility of the N-terminal region. Although many of the classical IP3K-A studies had focused on the its kinase activity, recent investigations have revealed that IP3K-A is also kinase-independently involved in the reorganization of the neuronal cytoskeleton through F-actin and microtubules bindings [6–8]. Interestingly, both of two cytoskeletal binding regions are located at the N-terminus of the enzyme [6,9], suggesting that the N-terminal regulatory domain may contribute to the spatiotemporal regulation of IP3K-A activity. In this sense, in vitro biochemical assays using purified recombinant IP3K-A are indispensable to investigate the physiological

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

roles of IP3K-A. To this end, a number of reports have explored methods for the expression and purification of recombinant IP3K-A tagged with various affinity tags, including glutathione S-transferase (GST), the NusA solubility-promoting tag, and polyhistidine (HIS) tags [6–10]. However, none of these endeavors was particularly effective, because the conventional expression systems using Escherichia coli resulted in low yields of full-length IP3K-A, poor enzyme solubility (even with the NusA system), and large amounts of truncated protein. To circumvent these problems, an alternative expression system named the cold-shock (pCold1) vector was used [11,12]. The pCold1 vector contains a cspA promoter that specifically stimulates gene expression via cold-shock induction, effectively minimizing the expression of bacterial host proteins [12]. Moreover, low-temperature incubation of E. coli gives rise to improve recombinant protein solubility and stability. The present study utilized the pCold1 system in an attempt to enhance the expression and solubility of IP3K-A, and compared the utility of the pCold1 system to that of three other bacterial expression systems: a HIS system, a GST system, and a maltose binding protein (MBP) system. The pCold1 system not only afforded increased yields of soluble IP3K-A, but its combined use with a C-terminal hexahistidine tag (C-HIS) also improved the purity of IP3K-A through the removal of truncated N-terminal fragments. These results suggest that our modified protocol for IP3K-A expression and purification will assist in the biochemical characterization of the enzyme, both structurally and functionally. Material and methods Vector construction To delete the sequence for the N-terminal hexahistidine tag (N-HIS) from the original pCold1-GST vector (available from TaKaRa Bio Inc.), deletion mutagenesis was performed by using a standard site directed mutagenesis with the following forward (FW) and reverse (RV) primers, as described previously [13]: Del N-HIS FW, 50 -ACC ATG AAT CAC AAA GTG CAT ATG GAG CTC ATG TCC-30 ; and Del N-HIS RV, 50 -GGA CAT GAG CTC CAT ATG CAC TTT GTG ATT CAT GGT-30 ). Next, a BamHI restriction site and the tobacco etch virus (TEV) cysteine protease cleavage site were introduced into the pCold1 vector by using general polymerase chain reaction (PCR) and the following forward primer: BamH1TEV-IP3K-A-FW, 50 -GAG GGA TCC GAA AAC CTG TAT TTT CAG GGC ATG ACC CTG CCC GGA CAC CCG ACG GGC-30 ; where the bold text denotes the BamHI restriction sequence, and the underlined text denotes the TEV cleavage sequence. In addition, a HindIII restriction site, a C-terminal hexahistidine tag (C-HIS), and a stop codon were introduced via the following reverse primer: IP3K-A-HIS-HindIII RV, 50 -GAC AAG CTT TCA ATG ATG ATG ATG ATG ATG TCT CTC AGC CAG GTT GGC CAA GAT GCC-30 ; where the bold text denotes the HindIII restriction sequence, the underlined text denotes the C-HIS sequence, and the italicized text denotes the stop codon. The IP3K-A insert was then amplified by PCR and digested with BamHI and HindIII. The digested insert and the pCold1 vector were ligated with a ligase mixture (TaKaRa), and the sequence of the sub-cloned vector (pCold1-GST-IP3K-A-HIS) was confirmed by a standard sequencing (Macrogen, South Korea). Primers used for generating 10xHIS-IP3K-A, MBP-IP3K-A, GST-IP3K-A and pCold1-GST-IP3K-A-HIS are summarized in Supplementary Table 1.

73

A single colony was selected the next day and incubated in LB broth (5 ml) containing ampicillin. Fully grown 1 ml of E. coli cultures were again inoculated into LB broth (200 ml), and the OD600 value was monitored for the optimized expression of recombinant IP3K-A. Under the present experimental conditions, an OD600 value of 0.6 was achieved after 2.5–3 h. At this time, the E. coli culture was rapidly transferred to a pre-cooled incubator at 15 °C for cold shock. After a 10-min incubation at 15 °C, isopropyl b-D-1-thiogalactopyranoside (IPTG, 0.1–1.0 mM) was added to 200-ml aliquots of the culture for 16–24 h to induce the expression of pCold1-GST-IP3K-A-HIS. Each overexpression protocol of pProEX-HTa-HIS-IP3K-A, pMAL-P2-IP3K-A, and pGEX-5X-1-IP3K-A were described in Supplementary material. C-HIS affinity purification of pCold1-GST-IP3K-A-HIS C-HIS affinity purification of pCold1-GST-IP3K-A-HIS was conducted based on interactions between the polyhistidine tag at the C-terminus of the enzyme and metal ions. An E. coli pellet acquired from 200 ml of culture was resuspended in HIS binding buffer (20 ml; 0.5 M NaCl, 20 mM imidazole, and 20 mM Tris–HCl, pH 7.9) containing a protease inhibitor cocktail (Roche). Sonication was performed with an ultrasonic cell disruptor (Branson Sonifier; amplitude: 35%, pulse-on: 10 s, pulse-off: 10 s, total time: 2 min). Whole cell lysates were cleared by centrifugation at 16,100g for 20 min. The soluble fraction was carefully transferred to a fresh 50 ml tube. Next, affinity columns (Bio-Rad, Cat# 731–1550) were packed with Ni2+-charged agarose resin (2 ml) (Elpis Biotech, Cat# EBE-1031). Two column volumes (CVs) of HIS binding buffer were loaded onto the column to allow for washout of the resin, and the cleared lysates were then loaded onto the agarose resin at a flow rate of approximately 2 ml/min. The flow-through was reloaded onto the column three times. The column was washed with 4–5 CVs of wash buffer (0.5 M NaCl, 140 mM imidazole, and 20 mM Tris–HCl, pH 7.9) to remove residual, nonspecifically bound proteins, and the protein concentration of the flow-through was monitored until the OD280 reached a value of zero. Recombinant pCold1-GST-IP3K-A-HIS IP3K-A was then eluted with elution buffer (5 ml; 0.5 M NaCl, 250 mM imidazole, and 20 mM Tris–HCl, pH 7.9), and the eluted protein solution (5 ml) was collected as five separate 1 ml fractions. GST affinity purification of pCold1-GST-IP3K-A-HIS Alternatively, pCold1-GST-IP3K-A-HIS was affinity purified based on interactions between the GST tag at the N-terminus of the enzyme and glutathione Sepharose beads. A bacterial pellet acquired from 200 ml of culture was resuspended in NETT buffer [20 ml; 50 mM Tris–HCl, pH 7.9, 100 mM NaCl, 5 mM EDTA, 0.5% Triton X-100, and 1 mM dithiothreitol (DTT)] containing a protease inhibitor cocktail (Roche). Sonication and lysate clearing were performed as described above. The cleared lysates were loaded onto a glutathione-agarose resin (2 ml) (Elpis Biotech, Cat# EBE-1041) at a flow rate of approximately 2 ml/min. The flow-through was reloaded onto the column three times. The column was washed with 4–5 CVs of NETT buffer to remove residual, nonspecifically bound proteins. Recombinant pCold1-GST-IP3K-A-HIS was then eluted with GST elution buffer (5 ml; 50 mM Tris–HCl, pH 7.9, 25 mM reduced glutathione, 0.1% Triton X-100, and 1 mM DTT), and the eluted protein solution (5 ml) was collected as five separate 1 ml fractions.

Overexpression of pCold1-GST-IP3K-A-HIS Calmodulin affinity purification of pCold1-GST-IP3K-A-HIS The pCold1-GST-IP3K-A-HIS vector was introduced into E. coli [BL21(DE3) strain] by heat shock at 42 °C for 50 s. Transformed E. coli cells were grown overnight on ampicillin-containing plates.

Finally, pCold1-GST-IP3K-A-HIS was affinity purified by virtue of its ability to bind to calmodulin. A bacterial pellet acquired from

74

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

200 ml of culture was resuspended in calmodulin binding buffer (20 ml; 50 mM Tris–HCl, pH 7.9, 100 mM NaCl, and 2 mM CaCl2) containing a protease inhibitor cocktail (Roche). Sonication and lysate clearing were performed as described above. The cleared lysates were loaded onto a calmodulin Sepharose resin (2 ml) (GE Healthcare, Cat# 71–7076-00) at a flow rate of approximately 2 ml/min. The flow-through was reloaded onto the column three times. The column was washed with 4–5 CVs of calmodulin binding buffer to remove residual, nonspecifically bound proteins. Recombinant pCold1-GST-IP3K-A-HIS was eluted with calmodulin elution buffer (5 ml; 50 mM Tris–HCl pH 7.9, 100 mM NaCl, and 2 mM EGTA), and the eluted protein solution (5 ml) was collected as five separate 1 ml fractions. Immunoblotting analysis After sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE), proteins were transferred onto a nitrocellulose membrane (Whatman Protran, Germany), and blocked with 5% skim milk in TBST (50 mM Tris–HCl pH 7.4, 150 mM NaCl, 0.1% Tween20) for 1 h. Then, the membrane was incubated with primary antibodies for 1 h at room temperature (RT) or overnight at 4 °C. After extensive washes 3 times, the membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 30 min. After further washes 3 times, the membrane was developed using the ECL system (Thermo Scientific Pierce). Antibodies Antibodies used in present study and antibody titers for Western blotting analysis are as followings; mouse anti-IP3K-A antibody (homemade, clone II4E; 1:1000), mouse N-terminal specific anti-IP3K-A antibody (homemade, clone VI5G; 1:1000), mouse anti-HIS antibody (Thermo scientific, Cat# MA1–21315; 1:2000), mouse anti-IgG HRP (Jackson Lab; 1:10,000). Performance and specificity of homemade antibodies was confirmed by our previous articles [6,7]. Kinase-Glo luminescent kinase assay A kinase-Glo luminescent kinase assay kit (Promega, Cat# V6711) was used to assess the kinase activity of purified GSTIP3K-A-HIS, as described in manufacture manual. Briefly, purified IP3K-A (5 lg) recombinant protein were incubated in following 100 ll buffer conditions (50 mM Tris–HCl pH 7.4, 20 mM MgSO4) and supplemented with 100 lM IP3 (Avanti, Cat# 850115P) and 10 lM ATP (Tocris, Cat# 3245) at 37 °C for 1 or 2 h. In control group, IP3 was replaced with 10 ll deionized water. Next, 50 ll of reaction mixtures are added into 50 ll of kinase-Glo reagent on the solid white 96 well plate. After further 10 min incubation at 37 °C, luminescence of each well was measured using the luminometer (Synergy 2 Multi-Mode Microplate Reader, BioTek). Microtubule binding assay A microtubule binding protein binding assay kit (Cytoskeleton, Cat# BK029) was used to assess the functional integrity of purified GST-IP3K-A-HIS, as described previously [6]. Briefly, pCold1-GSTIP3K-A-HIS was purified by the C-HIS affinity method, and a mixture of purified recombinant enzyme plus purified tubulin was then suspended in general tubulin buffer (Cytoskeleton, Cat# BST01–001; 80 mM PIPES buffer, pH 7.0, 2 mM MgCl2, and 0.5 mM EGTA). Next, a solution of 1 mM GTP (Cytoskeleton, Cat# BST06) plus 20 mM taxol (Cytoskeleton, Cat# TXD01) was added to the pCold1-GST-IP3K-AHIS/tubulin mixture to initiate microtubule polymerization, followed by incubation at room temperature for 30 min. Microtubule cushion

buffer (100 ll) (Cytoskeleton, Cat# BST05–001; general tubulin buffer in 50% (v/v) glycerol) was then placed into a centrifuge tube, and the pCold1-GST-IP3K-A-HIS/tubulin reaction mixture was layered on top of the cushion buffer and centrifuged at 100,000g at room temperature for 40 min. The supernatant was collected, and the pellet was resuspended in SDS sample buffer for analysis of SDS–PAGE. Results and discussion Enhanced expression and solubility of pCold1-GST-IP3K-A-HIS Affinity-tag proteins (e.g., GST, HIS, and MBP) generally enhance the stability and solubility of their target proteins in bacterial expression systems [13]. However, the use of GST and MBP has not been particularly effective in the case of IP3K-A, because both GST-IP3K-A and MBP-IP3K-A exhibit extensive fragmentation at the N-terminal region, poor solubility, and low expression levels in E. coli. To overcome these problems, we introduced the pCold1 vector, which permits the high-yield expression of target proteins under low-temperature conditions, with enhanced solubility. Moreover, the low-temperature incubation promotes the expression of recombinant target proteins, while simultaneously repressing the expression of bacterial host proteins [12]. The current study modified the original pCold1-GST vector to create the C-HIS affinity tag, with the goal of excluding N-terminal IP3K-A fragments during the purification scheme. To evaluate the performance of the pCold1-GST-IP3K-A-HIS vector, we compared the expression profiles and solubility of four different IP3K-A constructs with various affinity tags, as follows: N-HIS-IP3K-A (Fig. 1A), MBP-IP3K-A (Fig. 1B), GST-IP3K-A (Fig. 1C), and pCold1-GST-IP3K-A-HIS (Fig. 1D). After IPTG induction, total cell extracts and supernatant fractions were prepared from E. coli [BL21 (DE3) strain] for each expression vector and resolved by SDS–PAGE. Expression of each IP3K-A construct was evaluated in the absence or presence of IPTG after staining of the gel with Coomassie brilliant blue (CBB) (Fig. 2A). The yellow arrows denote polypeptide bands that were only observed following IPTG induction. Notably, only pCold1-GST-IP3K-A-HIS was observed with the unassisted eye in the supernatant portion (Fig. 2A, red arrowhead), suggesting that the pCold1-GST-IP3K-A-HIS vector system enhanced the solubility of recombinant IP3K-A. To investigate the identity and expression levels of recombinant IP3K-A, we performed Western blots immunoreacted with an IP3KA-specific monoclonal antibody (Fig. 2B). In the result of Western blot, we could confirm that all four vectors expressed full-length IP3K-A (Fig. 2B, yellow arrows) including partial fragments (Fig. 2B, red arrowheads). Intensities of the full-length bands were measured with ImageJ software (National Institutes of Health) and displayed as a histogram (Fig. 2C). High levels of full-length IP3K-A were observed in three constructs, MBP-IP3K-A, GST-IP3K-A, and pCold1-GST-IP3K-A-HIS (Fig. 2B and C). Among these three constructs, pCold1-GST-IP3K-A-HIS show the highest intensity of immunoreactivity, thus indicating the superiority of pCold1-GSTIP3K-A-HIS. Optimization of IPTG induction conditions for pCold1-GST-IP3K-A-HIS expression IPTG induction conditions for the pCold1-GST-IP3K-A-HIS vector were further optimized, as shown in Fig. 3. E. coli containing the vector were treated with various concentrations of IPTG (0.1, 0.5, or 1.0 mM), and the cultures (2 ml) were harvested 8, 16, 24, and 48 h later. Culture supernatants were resolved by SDS–PAGE. Yellow arrows denote the presumed full-length IP3K-A polypeptide band after CBB staining of the gel (Fig. 3A). The maximum yields

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

75

Fig. 1. Vector maps for IP3K-A constructs used in the present study. (A) The vector map of pProEX-HTa-HIS-IP3K-A. This vector consists of N-terminus 10xHIS and 1.4 kb IP3K-A full-sequence. The full-length sequence is 6072 base pairs. Full-sequence of IP3K-A was inserted into the vector by BamHI and HindIII restriction sites. (B) The vector map for pMAL-P2-IP3K-A consists of N-terminus MBP tag, Factor Xa cleavage site, and 1.4 kb IP3K-A full-sequence. The full-length vector contains 8079 base pairs. (C) The vector map of pGEX-5X-1-IP3K-A. This vector consists of 8079 base pairs. The GST tag and Factor Xa sites are located at the N-terminus of 1.4 kb IP3K-A sequence. (D) Schematic vector map of pCold1-GST-IP3K-A-HIS. This full-length vector contains 6489 base pairs. The GST tag, the human rhinovirus (HRV) 3C protease site, and the TEV protease cleavage site are located at the N-terminus of IP3K-A. The original N-terminus HIS tag of pCold1 vector was removed by the method of site-directed mutagenesis. Instead, the C-terminal HIS tag was newly introduced for our experimental purposes. (A–D) Each promoter, fused tag, and IP3K-A of four vectors was represented as a type of solid bar under each of the four vector maps. All vector maps in Fig. 1 were drawn and modified by SnapGene 2.2.1 and Adobe Illustrator CS5 software. (E) The schematic figure represents the structural domain and binding regions of IP3K-A including F-actin, calmodulin, and microtubules.

76

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

Fig. 2. The pCold1-GST-IP3K-A-HIS vector increases the expression and solubility of recombinant IP3K-A. (A) For the direct comparison of expression levels and solubility of various IP3K-A vectors, we evaluated four different expression vectors of IP3K-A as follows; N-HIS-IP3K-A, MBP-IP3K-A, and GST-IP3K-A, and pCold1-GST-IP3K-A-HIS. E. coli whole cell lysates (Tot.) and soluble supernatants (Sup.) from uninduced and IPTG-induced cultures were resolved by SDS–polyacrylamide gels and stained with CBB. The yellow arrows denote polypeptide bands that were only observed after IPTG induction, at a molecular weight close or identical to the expected molecular weight of recombinant full-length IP3K-A. A prominent band was observed in the soluble fraction for pCold1-GST-IP3K-A-HIS, but not for N-HIS-IP3K-A, MBP-IP3K-A, or GST-IP3K-A. (B) SDS–PAGE and Western blot analysis of whole cell lysates and soluble fractions with a primary monoclonal antibody specific for IP3K-A. The yellow arrows denote polypeptide bands of the correct size corresponding to the recombinant full-length IP3K-A constructs. Note that many fragments are apparent due to the instability of the IP3K-A N-terminal region. (C) The histogram displays the immunoreactive intensity of each band indicated with yellow arrows in Fig. 2B. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

of IP3K-A were observed with 0.5 and 1.0 mM IPTG and at least 16 h of induction time (Fig. 3A). We also determined the optimal bacterial concentration for IPTG induction by monitoring the OD600 value of the cultures. Transformed E. coli were grown until various OD600 values were reached (from 0.2 to 0.8, at an interval of 0.2) and then induced with 50 mM IPTG. Cultures with OD600 values of 0.6–0.8 provided the best performance in terms of both IP3K-A expression and solubility. The identity of the generated proteins was confirmed by Western blotting with the IP3K-A-specific monoclonal antibody (Fig. 2C) and quantified by using ImageJ software (Fig. 2D). Immunoreactive polypeptide bands with the appropriate size for fulllength recombinant IP3K-A predominated in the culture supernatant of the optimized E. coli samples, confirming that the pCold1GST-IP3K-A-HIS vector facilitated the expression and, especially, the solubility of recombinant IP3K-A. Purification of pCold1-GST-IP3K-A-HIS Based on the optimized induction conditions described above for pCold1-GST-IP3K-A-HIS (IPTG concentration: 0.5 mM, bacterial titer: OD600 0.8, induction time: 24 h), we set out to purify a recombinant IP3K-A protein. As shown in Fig. 2B, bacterial expression of pCold1-IP3K-A-HIS resulted in the generation of full-length

enzyme, in addition to an assortment of various-sized IP3K-A fragments. This protein fragmentation is likely due to the intrinsic instability of the N-terminal region. To purify only the full-length construct, we first selected a C-HIS-mediated affinity purification system, excluding the GST and calmodulin affinity systems discussed below. The general approach for C-HIS-mediated affinity purification is illustrated in Supplementary Fig. S1 and utilized an affinity column packed with Ni2+-charged agarose resin. In a preliminary affinity purification experiment, a major polypeptide band with a molecular weight of about 75 kDa (yellow arrow), corresponding to fulllength GST-IP3K-A-HIS, was eluted from the column with 250 mM imidazole (Fig. 4A). Unfortunately, several smaller bands with molecular weights of about 25–40 kDa were also eluted from the Ni2+-charged agarose resin (red arrowheads), although these fragments occurred with a lower frequency. In previous study, it was also reported that IP3K-A is fragmented by some proteases such as calpain [14]. To reveal the identification of fragments, we performed Western blot analysis using both HIS and N-terminal specific IP3K-A antibody. In the result of Western blot, most of all fragments had no C-HIS affinity tags while large portions of fragments were blotted by N-terminal specific IP3K-A antibody (Supplementary Fig. S2). Therefore we could conclude that most of these fragments were derived from the N-terminal portion of

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

77

Fig. 4. Optimization of C-HIS affinity purification of pCold1-GST-IP3K-A-HIS. (A) Preliminary purification of pCold1-GST-IP3K-A-HIS based on C-HIS binding to the Ni2+-charged agarose resin. (B) Optimization of the imidazole concentration for the elution of full-length pCold1-GST-IP3K-A-HIS. (A and B) Yellow arrows denote fulllength constructs, while red arrowheads indicate partial fragments. SDS–polyacrylamide gel was stained with CBB. Used abbreviations are as followings; Sup., soluble culture supernatant fraction; FT, flow-through fraction; Wash, washout fraction; Elu, eluted fraction. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Optimization of IPTG induction conditions for pCold1-GST-IP3K-A-HIS expression. (A) SDS–polyacrylamide gels stained with CBB are showing expression profile of pCold1-GST-IP3K-A-HIS under various experimental conditions with different combinations of IPTG concentrations (0.1, 0.5, or 1.0 mM) and induction times (8, 16, 24, or 48 h). The yellow arrows denote predominant polypeptide bands that were only apparent after IPTG induction. (B) SDS–polyacrylamide gel stained with CBB showing expression levels of pCold1-GST-IP3K-A-HIS with IPTG at a fixed concentration of 0.5 mM and various bacterial titers (OD600 values, 0.2, 0.4, 0.6, and 0.8). The yellow arrows denote predominant polypeptide bands after IPTG induction. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

IP3K-A. However, the question then arises as to how these untagged proteins could be isolated by C-HIS affinity purification. In this regard, IP3K-A can reportedly dimerize through its N-terminal region [15]. Accordingly, we conjectured that the N-terminal fragments observed in this study formed heterodimers with fulllength GST-IP3K-A-HIS and were thereby retained on the Ni2+charged resin with the assistance of full-length GST-IP3K-A-HIS.

Therefore, we then focused on the affinity difference between homodimers (two full-length GST-IP3K-A-HIS) and heterodimers (one full-length GST-IP3K-A-HIS and one partial N-terminal fragment of IP3K-A), because a heterodimer contains only a single C-HIS tag, whereas a homodimer contains two C-HIS tags. Moreover, we hypothesized that a specific imidazole concentration might separate the putative heterodimers from the homodimers due to a slight binding difference by the number of C-HIS tag. To test this hypothesis, we eluted pCold1-GST-IP3K-A-HIS from the Ni2+-charged resin by using a gradient of imidazole concentrations that ranged from 40 to 200 mM, at intervals of 20 mM. SDS–PAGE revealed that polypeptide bands corresponding to fragmented IP3K-A (red arrowheads) were eluted in the 140–160-mM imidazole fraction, whereas large portions of full-length IP3K-A (yellow arrow) were first eluted in the 180-mM imidazole fraction (Fig. 4B). Thus, this small difference in imidazole elution concentrations will be quite effective for the separation of full-length IP3K-A from its partial fragments. Comparison of C-HIS, GST, and calmodulin affinity purification systems Next, we evaluated the efficacy of C-HIS affinity purification of pCold1-GST-IP3K-A-HIS (as assessed by yield and solubility of full-length protein) relative to other established affinity purification systems, i.e., GST and calmodulin purification systems [6,7,14]. For a fair comparison, the conditions for each purification system were equalized to the greatest extent possible, with the exception of specific buffer constituents, and the washout step was performed until no proteins appeared in the flow-through fraction. Samples from each purification step were kept for analysis

78

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

by SDS–PAGE. As expected, the final eluted fractions in the GST and calmodulin affinity purification systems were enriched in partial IP3K-A fragments (Fig. 5A and B), because the GST affinity tag and the calmodulin binding site are located at the N-terminus and within the N-terminal region (amino acids 156–189 out of a total of 459 amino acids) of the enzyme, respectively. On the other hand, the C-HIS affinity purification system, together with 180 mM imidazole as the eluting buffer, exhibited only slight amounts of contaminating partial fragments (which likely stemmed from the effect of heterodimerization, as discussed above) (Fig. 5C).

The final eluted fractions from the three affinity purification systems are shown side-by-side in Fig. 4D. Quantitative analysis of IP3KA generation was conducted by measuring band intensities along the vertical axis of the lanes with Image J software (Fig. 5E–G). The amount of full-length IP3K-A was calculated as a percentage of measured band intensity (area of full-length peak divided by total area of peaks). The C-HIS system yielded the highest amount of full-length enzyme (87.0%) among the three purification systems (Fig. 5H), indicating that the newly introduced C-HIS/pCold vector system improves the purity of recombinant IP3K-A compared with the

Fig. 5. Comparison of three affinity purification systems for pCold1-GST-IP3K-A-HIS (A-C) SDS–PAGE analysis of pCold1-GST-IP3K-A-HIS affinity purification conducted by using a GST affinity system with a glutathione-bound Sepharose resin (A), a calmodulin binding protein affinity system with a calmodulin-bound Sepharose resin (B), and a C-HIS affinity system with a Ni2+-charged agarose resin (C). (D) Final eluted fractions for the three purification systems are shown. Each gel image corresponds to the reddotted rectangles in (A–C). Used abbreviations are as followings; Sup., soluble culture supernatant fraction; FT, flow-through fraction; Wash, washout fraction; Elu, eluted fraction. (E–G) Band intensity profiles of eluted fractions acquired from the GST (E, peak of full-length: 1–4), calmodulin (F, peak of full-length: 2–4), and C-HIS purification schemes (G, peak of full-length: 3–3). (H) Histogram showing the purity of full-length pCold1-GST-IP3K-A-HIS for each affinity system (GST: 31.8%, calmodulin (CaM): 63.8%, C-HIS: 87.0%). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

conventional GST (31.8%) and calmodulin (63.8%) affinity systems reported previously [6,7,14]. Functional confirmation of purified pCold1-GST-IP3K-A-HIS by kinase activity assay and microtubule binding assay It is important to select suitable screening tests to confirm the functional integrity of a purified recombinant protein. To do this, we focused two intrinsic properties of IP3K-A, its kinase activity and microtubule binding affinity. Kinase activity assay is most straightforward way to confirm functional integrity of kinases. To measure kinase activity of our recombinant IP3K-A, we introduced an indirect method of Kinase-Glo Luminescent Kinase Assay (Promega, Cat# V6711) which quantitate the amount of ATP remaining in reactions after kinase reactions. This assay is based on the luciferase reaction which produces one photon of light consuming one molecule of ATP [16]. If IP3K-A converts IP3 into IP4 specifically, IP3K-A will consume ATP only in the presence of IP3, its substrate of kinase reactions. Therefore luminescence of this assay is inversely co-related to kinase activity of IP3K-A. Fig. 6A represents general reaction equations of this kinase activity assay. Purified pCold1-GST-IP3K-A-HIS was incubated with 100 lM IP3 and 10 lM ATP. After 1 or 2 h incubation, we measured the amount of ATP remaining in 4 independent wells of each groups. Expectedly ATP level was greatly reduced in the presence of IP3 (Fig. 6B and Supplementary Table 2; ATP reduction percentage for 1- and 2-h reaction times was 58.4 ± 1.5% and 41.4 ± 0.6%, respectively; mean ± SEM, Student’s ttest, ⁄⁄⁄p = 5.06E-09 for 1 h incubation, and ⁄⁄⁄p = 1.67E-10 for 2 h incubation), suggesting that a catalytic function of IP3K-A is working properly. As noted above, the N-terminal region of IP3K-A is notorious for its structural instability and flexibility. For this reason, two

79

crystallographic studies of IP3K-A were restricted to only the catalytic domain at the C-terminal region of the enzyme [4,5]. However, recent studies have revealed the importance of the N-terminus region, which regulates various key properties of IP3K-A, including its subcellular localization and ability to reorganize the neuronal cytoskeleton [6–8]. Therefore, a confirmation test for IP3K-A function must also embrace the N-terminal region, despite it being the most failure-prone part of the protein. In this sense, the microtubule binding assay is perhaps the best choice, because the entire N-terminus region of IP3K-A is strongly involved in the enzyme’s binding to microtubules [6]. Therefore, because any partial defects in the N-terminal region can also affect its binding affinity for microtubules, this assay can effectively evaluate both the structural and functional integrity of IP3K-A. The microtubule binding assay is based on the fact that microtubule binding proteins co-sediment with polymerized microtubules following high-velocity centrifugation (Supplementary Fig. S3). We found that purified pCold1-GST-IP3K-A co-sedimented in the pellet fraction in the presence of polymerized microtubules, but not in their absence (Fig. 6C). These results suggest that purified pCold1-GST-IP3K-A retains its functional integrity. To calculate the efficiency of microtubule-dependent pelleting, we employed the following equation: [efficiency of microtubuledependent pelleting = band intensity of pCold1-GST-IP3K-A in the pellet/(band intensity of pCold1-GST-IP3K-A in the supernatant) + (band intensity of pCold1-GST-IP3K-A in the pellet)], as described previously [6]. Three independent experiments were performed. We obtained statistically significant results (control: 13.24 ± 1.16%, experimental group: 81.17 ± 10.9%; mean ± SEM, Student’s t-test, ⁄⁄⁄p = 0.000427) (Fig. 6D). Taken together, these observations demonstrate that pCold1-GST-IP3K-A-HIS expressed and purified by our modified protocol is structurally and functionally intact.

Fig. 6. Functional confirmation of purified pCold1-GST-IP3K-A-HIS via kinase activity assay and the microtubule binding assay. (A) General reaction equations of luminescent kinase assay. ATP consuming by IP3K-A and phosphorylation of IP3 induces low level of luminescence. (B) The histogram depicts with the result of kinase activity assay. Average percentage of ATP reduction for 1- and 2-h reaction times was 58.4 ± 1.5% (mean ± SEM, Student’s t-test, ⁄⁄⁄p = 5.06E-09) and 41.4 ± 0.6% (mean ± SEM, Student’s ttest, ⁄⁄⁄p = 1.67E-10), respectively. (C) SDS–polyacrylamide gels stained with CBB revealed that pCold1-GST-IP3K-A-HIS was precipitated in the presence of polymerized microtubules, but not in their absence. (D) Quantitative analysis of band intensities, as assessed by the following equation: [efficiency of microtubule-dependent pelleting = band intensity of pCold1-GST-IP3K-A-HIS in the pellet/(band intensity of pCold1-GST-IP3K-A-HIS in the supernatant) + (band intensity of pCold1-GST-IP3K-A-HIS in the pellet)], where a higher percentage indicates stronger binding to microtubules.

80

D. Lee et al. / Protein Expression and Purification 97 (2014) 72–80

Conclusion This study presents a modified protocol for the efficient overexpression and purification of recombinant IP3K-A in E. coli by using the pCold1 vector and a C-terminal hexahistidine affinity tag, C-HIS. Our improved method is likely to be helpful for elucidating the structural and functional profiles of IP3K-A. Acknowledgments This study was supported by the Bio & Medical Technology Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (MEST) (2012M3A9B6055378 to H. Kim) and the Brain Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2013-035006 to H. Kim). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.pep.2014.02.006. References [1] M. Yamada, A. Kakita, M. Mizuguchi, S.G. Rhee, S.U. Kim, F. Ikuta, Specific expression of inositol 1,4,5-trisphosphate 3-kinase in dendritic spines, Brain Res. 606 (1993) 335–340. [2] I.H. Kim, S.K. Park, W. Sun, Y. Kang, H.T. Kim, H. Kim, Spatial learning enhances the expression of inositol 1,4,5-trisphosphate 3-kinase A in the hippocampal formation of rat, Brain Res. Mol. Brain Res. 124 (2004) 12–19. [3] M.J. Schell, Inositol trisphosphate 3-kinases: focus on immune and neuronal signaling, Cell. Mol. Life Sci. 67 (2010) 1755–1778.

[4] G.J. Miller, J.H. Hurley, Crystal structure of the catalytic core of inositol 1,4,5trisphosphate 3-kinase, Mol. Cell 15 (2004) 703–711. [5] B. Gonzalez, M.J. Schell, A.J. Letcher, D.B. Veprintsev, R.F. Irvine, R.L. Williams, Structure of a human inositol 1,4,5-trisphosphate 3-kinase: substrate binding reveals why it is not a phosphoinositide 3-kinase, Mol. Cell 15 (2004) 689–701. [6] D. Lee, H.W. Lee, S. Hong, B.I. Choi, H.W. Kim, S.B. Han, I.H. Kim, J.Y. Bae, Y.C. Bae, I.J. Rhyu, W. Sun, H. Kim, Inositol 1,4,5-trisphosphate 3-kinase A is a novel microtubule-associated protein: PKA-dependent phosphoregulation of microtubule binding affinity, J. Biol. Chem. 287 (2012) 15981–15995. [7] I.H. Kim, S.K. Park, S.T. Hong, Y.S. Jo, E.J. Kim, E.H. Park, S.B. Han, H.S. Shin, W. Sun, H.T. Kim, S.H. Soderling, H. Kim, Inositol 1,4,5-trisphosphate 3-kinase a functions as a scaffold for synaptic Rac signaling, J. Neurosci. 29 (2009) 14039– 14049. [8] S. Windhorst, C. Blechner, H.Y. Lin, C. Elling, M. Nalaskowski, T. Kirchberger, A.H. Guse, G.W. Mayr, Ins(1,4,5)P3 3-kinase-A overexpression induces cytoskeletal reorganization via a kinase-independent mechanism, Biochem. J. 414 (2008) 407–417. [9] M.J. Schell, C. Erneux, R.F. Irvine, Inositol 1,4,5-trisphosphate 3-kinase A associates with F-actin and dendritic spines via its N terminus, J. Biol. Chem. 276 (2001) 37537–37546. [10] P.J. Woodring, J.C. Garrison, Expression, purification, and regulation of two isoforms of the inositol 1,4,5-trisphosphate 3-kinase, J. Biol. Chem. 272 (1997) 30447–30454. [11] K. Hayashi, C. Kojima, PCold-GST vector: a novel cold-shock vector containing GST tag for soluble protein production, Protein Expr. Purif. 62 (2008) 120–127. [12] G. Qing, L.C. Ma, A. Khorchid, G.V. Swapna, T.K. Mal, M.M. Takayama, B. Xia, S. Phadtare, H. Ke, T. Acton, G.T. Montelione, M. Ikura, M. Inouye, Cold-shock induced high-yield protein production in Escherichia coli, Nat. Biotechnol. 22 (2004) 877–882. [13] A. Malhotra, Tagging for protein expression, Methods Enzymol. 463 (2009) 239–258. [14] S.Y. Lee, S.S. Sim, J.W. Kim, K.H. Moon, J.H. Kim, S.G. Rhee, Purification and properties of D-myo-inositol 1,4,5-trisphosphate 3-kinase from rat brain. Susceptibility to calpain, J. Biol. Chem. 265 (1990) 9434–9440. [15] H.W. Johnson, M.J. Schell, Neuronal IP3 3-kinase is an F-actin-bundling protein: role in dendritic targeting and regulation of spine morphology, Mol. Biol. Cell 20 (2009) 5166–5180. [16] M. Koresawa, T. Okabe, High-throughput screening with quantitation of ATP consumption: a universal non-radioisotope, homogeneous assay for protein kinase, Assay Drug Dev. Technol. 2 (2004) 153–160.