Heterologous expression and purification of an ... - Wiley Online Library

Heterologous expression and purification of an active human TRPV3 ion channel € m1,†, Stefan Kol1,*, Christian Braun2, Gerhard Thiel2, Declan A. Doyle3, Michael Sundstro 4 4 Pontus Gourdon and Poul Nissen 1 2 3 4

Protein Function and Interactions, Novo Nordisk Foundation Center for Protein Research, University of Copenhagen, Denmark Plant Membrane Biophysics, Technische Universit€at Darmstadt, Germany Centre for Biological Sciences, University of Southampton, UK Centre for Membrane Pumps in Cells and Disease, Danish National Research Foundation, Aarhus University, Denmark

Keywords ion channels; membrane proteins; protein expression; transient receptor potential channels; TRPV3 Correspondence S. Kol, Novo Nordisk Foundation Center for Biosustainability, Technical University of 6, DK-2970 Hørsholm, Denmark, Kogle Alle Denmark Fax: +45 4525 8001 Tel.: +45 2465 4189 E-mail: [email protected] Present address *Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Hørsholm, Denmark †Karolinska Development, Solna, Sweden

The transient receptor potential vanilloid 3 (TRPV3) cation channel is widely expressed in human tissues and has been shown to be activated by mild temperatures or chemical ligands. In spite of great progress in the TRP-channel characterization, very little is known about their structure and interactions with other proteins at the atomic level. This is mainly caused by difficulties in obtaining functionally active samples of high homogeneity. Here, we report on the high-level Escherichia coli expression of the human TRPV3 channel, for which no structural information has been reported to date. We selected a suitable detergent and buffer system using analytical size-exclusion chromatography and a thermal stability assay. We demonstrate that the recombinant purified protein contains high a-helical content and migrates as dimers and tetramers on native PAGE. Furthermore, the purified channel also retains its current inducing activity, as shown by electrophysiology experiments. The ability to produce the TRPV3 channel heterologously will aid future functional and structural studies. Structured digital abstract

(Received 17 June 2013, revised 6 September 2013, accepted 9 September 2013)

TRPV3 and TRPV3 bind by molecular sieving (1, 2) TRPV3 and TRPV3 bind by blue native page (1, 2, 3)

doi:10.1111/febs.12520

Introduction Transient receptor potential (TRP) proteins are cationselective channels that become permeable in response to a wide variety of physical, chemical and thermal signals [1,2]. Members of this family are conserved in yeast, invertebrates and vertebrates. TRP channels allow multicellular organisms to sense their surroundings and

single cells to respond to changes in their local environment. In humans, TRP channels can be grouped into six subfamilies on the basis of amino acid sequence homology: TRPC (canonical), TRPV (vanilloid), TRPM (melastatin), TRPA (ankyrin), TRPP (polycystin) and TRPML (mucolipin). All TRP channels are

Abbreviations 2-APB, 2-aminoethoxydiphenyl borate; CMC, critical micelle concentration; CPM, N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl] maleimide; DDDMG, N-dodecyl-N,N-dimethylglycine; DDM, n-dodecyl b-D-maltopyranoside; DDMAB, N-dodecyl-N,N-(dimethylammonio) butyrate; DDMG, N-decyl-N,N-dimethylglycine; FC, Fos-choline; GFP, green fluorescent protein; IMAC, immobilized metal-affinity chromatography; LDAO, lauryldimethylamine-N-oxide; LFC12, lysoFos-choline-12; SEC, size-exclusion chromatography; TRP, transient receptor potential; TRPV3, transient receptor potential vanilloid subtype 3.

6010

FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

S. Kol et al.

predicted to contain six transmembrane segments and assemble into tetrameric complexes. The N- and C-termini of TRP channels are located in the cytoplasm and, depending on the TRP-channel subfamily, consist of various functional domains such as ankyrin repeats, Ca2+-sensing EF hands, phosphorylation sites, calmodulin-binding sites and a so-called TRP box [3]. Because TRP channels are involved in many different pathways, their malfunctioning and dysregulation can lead to a wide array of human diseases, like neuropathic pain, cancer, neurological disorders, cardiovascular disorders and polycystic kidney disease [4]. TRP vanilloid 3 (TRPV3) was originally identified in database searches for TRP-related expressed sequence tags from a human testis library [5] and homology searches of databases for TRPV1, the founding member of the vanilloid subfamily of receptors [6–8]. Human TRPV3 is a 91 kDa integral plasma membrane protein with six transmembrane helices (TM1–6) formed by residues 440–670 which encompasses a pore region consisting of the ‘P-loop’ and TM5—6. The N-terminus contains three ankyrin repeats, whereas the C-terminus contains a consensus PDZ-binding motif (residues ETSV) and the TRP box, a common TRP-channel family sequence feature (residues IWRLQR). TRPV3 is highly expressed in the skin and is activated by temperatures in the range of 31–39 °C and by the chemicals camphor and 2-aminoethoxydiphenyl borate (2-APB) [9,10]. The TRPV3 channel, along with other TRPV channels, may play an important role in chronic pain and, therefore, is receiving increasing attention as a potential therapeutic candidate for the treatment of chronic pain [4]. Structures of isolated TRP-channel domains have been determined by X-ray crystallography or NMR [11,12] and low-resolution electron microscopy structures of full-length proteins have been obtained [13–16]. However, isolated domain structures provide little information about how these domains interact and thereby modulate the pore function. Furthermore, the electron microscopy structures have been inconclusive and are difficult to interpret functionally. Structural information of TRPV3 could be used to understand the atomic details of ligand binding and temperature sensitivity, and might help in designing new drug compounds. The first step toward getting a structure of TRPV3 is to obtain sufficient amounts of highly pure and stable protein for structural analysis. This would entail producing milligram amounts of TRPV3. Highyield heterologous protein production has been a major limiting step in the structural characterization of membrane proteins and these problems have been especially daunting for eukaryotic ion channels. Yeast or insect FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

Purification of human TRPV3

cells have been used to produce TRP channels, but such expression systems are often not adequate to generate the tens of milligrams (or more) that are typically needed to optimize the crystallization conditions producing diffraction-quality crystals. Production of eukaryotic membrane proteins in Escherichia coli normally requires considerable time and effort to achieve the high levels needed for purification, while retaining function [17]. Here, we present our efforts to optimize protein expression, solubilization and purification of human TRPV3 produced in E. coli. We have cloned 19 members of the human TRP family covering all subfamilies and tested their expression. Full-length human TRPV3 expressed particularly well and a green fluorescent protein (GFP) fusion facilitated detergent screening to improve sample homogeneity. Furthermore, because the purified channel is suitable for subsequent structural studies, we conducted a buffer screen to enhance chromatographic behavior and the likeliness of obtaining crystallization hits. A three-step purification procedure resulted in monodisperse and highly pure protein. CD and native PAGE were used to show that TRPV3 adopts an a-helical conformation and forms dimers and tetramers, respectively. Importantly, the structural integrity and function of the purified protein was confirmed by electrophysiology experiments.

Results High-throughput cloning and expression Initially, our goal was to clone all 27 human TRP channels. However, some expression constructs could not be completed due to problems obtaining the PCR insert or with subsequent cloning steps. The resulting 19 constructs, all with engineered hexa-histidine tags in the N-terminus, were transformed into E. coli strain BL21 (DE3)-R3-pRARE2 and analyzed in small-scale expression screening experiments in 96-well format. After Ni2+ chelating affinity chromatography, clear protein bands were observed for TRPV3 and TRPV5 (Fig. 1A, lane 16 and 18) as further supported through immunoblotting using a-his antibody (Fig. 1B). Although TRPV5 is clearly visible using Coomassie Brilliant Blue staining, it reacts poorly to the a-his antibody. The molecular size of the overexpressed TRPV3 and TRPV5 proteins was estimated to be ~ 75 and ~ 63 kDa, respectively. Several other low and high molecular mass protein bands were detected by immunoblotting using a-his antibody (Fig. 1B, lane 16). Later experiments showed that n-dodecyl b-D-maltopyranoside (DDM) behaves very poorly at keeping 6011

S. Kol et al.


A

A

B

C

D

B

Fig. 1. Human TRPV3 and TRPV5 are highly expressed in E. coli. (A) Cell lysates of E. coli containing TRP-channel expression constructs analyzed by SDS/PAGE. Gels were stained with (A) Coomassie Brilliant Blue dye or (B) subjected to western blotting using anti-(His tag) Ig. Particularly well-expressed proteins are denoted by asterisks. The observed molecular mass of the TRPV3 (93.1 kDa) and TRPV5 (85.1 kDa) proteins was ~ 75 and ~ 63 kDa, respectively.

TRPV3 stably in solution and the extra bands are likely caused by aggregation and degradation events. To aid subsequent characterization, a C-terminal GFP fusion was created of both TRPV3 and TRPV5 in the pWaldo–GFPe vector [18]. Unfortunately, the cells harboring the TRPV5–GFP fusion showed a decrease in optical density after induction and displayed no GFP fluorescence. By contrast, the cells harboring the TRPV3–GFP fusion grew normally and the cell pellet was fluorescent upon visual inspection (not shown). Purified inner membranes (outer membrane removed through sucrose-gradient ultracentrifugation) of induced E. coli cells harboring the TRPV3 or TRPV3–GFP constructs revealed expressed bands estimated to be ~ 75 and ~ 98 kDa, respectively (Fig. 2A). Both bands are detected by an anti-(His tag) IgG1 (Fig. 2B) and a polyclonal anti-TRPV3 IgG (Fig. 2C). A single fluorescent band was observed in the membranes expressing TRPV3–GFP (Fig. 2D). In addition, the identity of recombinant TRPV3 was verified by MALDI–TOF MS. Because proper folding of GFP fused to the C-terminus of a target protein depends on the correct folding of the latter, fluorescence will only be detected when both GFP and the fusion protein are folded [18,19]. With the observations that TRPV3 is highly overproduced and correctly localized to the inner membrane of E. coli, we decided to focus our attention on the purification and characterization of TRPV3. 6012

Fig. 2. TRPV3 and TRPV3–GFP locate to the inner membrane of E. coli. Inner membrane fractions of E. coli expressing TRPV3 and TRPV3–GFP were analyzed by SDS/PAGE and the gel was stained with (A) Coomassie Brilliant Blue dye, (B) subjected to western blotting using anti-(His tag) IgG1 or (C) subjected to western blotting using anti-TRPV3 IgG. (D) In gel GFP fluorescence of E. coli inner membrane fractions. TRPV3 and TRPV3–GFP are denoted by an arrow or an asterisk, respectively.

Detergent screening The choice of detergent that is used to solubilize and keep the protein in solution is one of the most important parameters in membrane protein purification and stabilization. The popular detergents DDM and N-octyl-b-Dglucoside and octaethylene glycol monododecyl ether were previously tested and found to be unsuitable for both solubilization and purification of TRPV3 (data not shown). Lauryldimethylamine-N-oxide (LDAO) was able to solubilize TRPV3 efficiently, but was poor at keeping it in a monodisperse state. We proceeded with the selection of a suitable detergent aided by the TRPV3–GFP fusion protein using a detergent screen from Hampton Research. The screen contains many detergents used in the crystallization of membrane proteins including ionic detergents, nonionic detergents, zwitterionic detergents, nondetergent sulfobetaines and synthetic lipids. Although the critical micelle concentration (CMC) is low for some detergents in this screen, it should be a good indicator of solubilization efficiency. Detergent solubilization of inner membranes containing TRPV3–GFP was performed in 96-well plates. After centrifugation, the residual fluorescence of the supernatant was measured and calculated as a percentage of the fluorescence of the starting material (Fig. 3A). In general, a value > 80% yielded a clear band on a Coomassie Brilliant Blue stained SDS/PAGE gel (not shown). We picked the detergents that yielded high fluorescence recovery, a clear SDS/PAGE band, and were not considered to be too strong or prohibitively expensive. Selected were the zwitterionic detergents N-dodecyl-N,N-dimethylglycine (DDDMG), N-decyl-N, FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

S. Kol et al.


A

B

Fig. 3. Fos-choline detergents are suitable to solubilize and purify TRPV3. (A) Inner membrane fractions of E. coli expressing TRPV3–GFP were solubilized with Hampton Research detergent additive screen HR2-406. GFP fluorescence of the supernatant is expressed as a percentage of input material. (B) Analytical gel-filtration experiments of His tag purified TRPV3– GFP on a Superose 6 SEC column using the detergents LDAO, FC10, FC11, FC12, LFC12, DDDMG, DDMG and DDMAB. (C) Unfolding curves measured for TRPV3 (continuous line) and TRPV3–GFP (dotted line) in the presence of FC10 (brown), FC11 (green) and FC12 (blue) at 40 °C for 175 min.

C

N-dimethylglycine (DDMG), Fos-choline-10 (FC10), Fos-choline-12 (FC12), N-dodecyl-N,N-(dimethylammonio)butyrate (DDMAB) and LysoFos-choline-12 (LFC12). Based on earlier results (see above) and to complete the range of Fos-choline hydrocarbon chain lengths, the detergents LDAO and Fos-choline-11 (FC11) were added to this list. Next, we performed analytical gel-filtration experiments to assess monodispersity in the presence of the selected detergents. TRPV3–GFP was solubilized from the membrane environment, purified by immobilized metal-affinity chromatography (IMAC) and loaded onto a Superose 6 size-exclusion chromatography (SEC) column. Although the DDMG, DDDMG, LFC12, LDAO detergents were very efficient at extracting TRPV3–GFP from the membrane, they performed poorly at keeping it in a monodisperse state FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

(Fig. 3B). By contrast, Fos-choline detergents and DDMAB improved the height and resolution of the TRPV3–GFP peak. Because Fos-choline detergents have been used frequently in the purification of TRP channels previously [13,20,21], we decided to assess their stabilization properties on TRPV3 by further means. We measured the stability of both TRPV3 and TRPV3–GFP in FC10, FC11 and FC12 using a fluorescence-based unfolding assay [22,23]. In this assay, the dye N-[4-(7-diethylamino-4-methyl-3-coumarinyl) phenyl]maleimide (CPM) becomes fluorescent upon reacting with free sulfhydryl groups, which become accessible as the protein unfolds. Because cysteines are frequently located at helix–helix interaction sites [24], cysteine accessibility is a good measure of protein unfolding. As shown in Fig. 3C, slow unfolding 6013

S. Kol et al.


of TRPV3 and TRPV3–GFP is observed in the presence of Fos-choline detergents. The stability of TRPV3 is improved by increasing hydrocarbon chain length, where FC12 shows the highest stability with a half-life of ~ 63 min. The presence of the GFP moiety seems to confer some stability on TRPV3, because the rate of unfolding decreases to a half-life of ~ 111 min. Because the detergent FC12 performs very well in the solubilization of TRPV3 and in the analysis by analytical gel filtration and CPM unfolding assay, we decided to use it in the purification of TRPV3. Buffer screening To aid analysis and future functional and crystallographic studies, we optimized several buffer conditions. We assessed the chromatographic behavior of TRPV3 at a range of pH, with different salts and salt concentrations, and with various additives and additive concentrations. Results were compared with our standard gel-filtration buffer (20 mM Hepes–KOH pH 7.5, 500 mM NaCl, 10% glycerol, 4.5 mM FC12) and any improvements in monodispersity were evaluated by assessing the height of the TRPV3 peak and how well TRPV3 is resolved from any contaminations. The optimal pH was found to be 7.5 (Fig. 4, left); other pH values led to a decrease in the height of the TRPV3 peak. Next, we assessed various concentra-

tions of sodium and potassium chloride and found that 150 mM potassium chloride performed as well as 500 mM sodium chloride in increasing the TRPV3 peak and in resolving TRPV3 from any low molecular mass contaminants (Fig. 4, middle). Other concentrations of both sodium chloride and potassium chloride decreased the TRPV3 peak. Because a high sodium chloride concentration may complicate any subsequent crystallization trials, we decided to use 150 mM potassium chloride in future experiments. Addition of CaCl2 (5 mM) did not lead to an improvement in chromatographic behavior (data not shown). Last, we addressed the influence of different concentrations of the additives glycerol, sucrose and sorbitol (Fig. 4, right). Although the peak height in 10% glycerol and sorbitol is similar, sorbitol performed slightly better at resolving TRPV3 from low molecular mass contaminants. Other concentrations of glycerol, sucrose and sorbitol were not as efficient at improving the SEC-profile TRPV3. It was decided to perform quantitative size-exclusion purifications in 20 mM Hepes–KOH pH 7.5, 150 mM KCl, 10% sorbitol and 4.5 mM FC12. Purification of TRPV3 and TRPV3–GFP Total membranes of E. coli expressing TRPV3 and TRPV3–GFP were solubilized using FC12 and purified by IMAC. The main elution fractions were pooled and

Fig. 4. Buffer screening improves the chromatographic behavior of TRPV3. Following IMAC, the behavior of TRPV3 was assessed by analytical gel filtration in the presence of different buffers using a Superose 6 column. Scouting was performed by screening different pH values (left), different concentrations of sodium and potassium chloride (middle) and different concentrations of glycerol, sucrose and sorbitol (right).

6014


S. Kol et al.

subjected to SEC. Superose 6 and Superdex 200 SEC columns were employed in tandem to separate the peak of interest from high and low molecular mass contaminants, respectively. Only the peak containing TRPV3 or TRPV3–GFP that eluted from the Superose 6 column was collected and loaded onto the Superdex 200 column. TRPV3–GFP elutes from a Superdex 200 column as a single peak flanked by both a high and a low molecular mass shoulder (Fig. 5A). GFP fluorescence is, however, predominantly detected in the main peak and not in the shoulders (Fig. 5B). The shoulders therefore likely represent contaminants or aggregated material. SDS/PAGE analysis of TRPV3–GFP in FC12 (Fig. 5, inset) after gel-filtration indicates a high purity, although some contaminations or truncations and a double band of TRPV3–GFP are observed. Non-GFP-tagged TRPV3 is also eluted from a Superdex 200 column as a sharp, single peak (Fig. 5C), which yields homogenous material as judged by SDS/ PAGE (Fig. 5, inset). Although a minor high molecular mass shoulder could be observed (see below), there were fewer contaminations compared with TRPV3– GFP. Because GFP may influence the functionality of TRPV3 and TRPV3 displays better chromatographic behavior, it was decided to conduct further studies with the homogenous preparations of nonfused TRPV3 in FC12. The total yield was 1.2–1.5 mg of purified protein per liter of E. coli culture.


A

B

C

Characterization of purified TRPV3 All TRP channels are thought to form tetrameric assemblies [25]. To establish whether purified TRPV3 is correctly folded and form homo-oligomers, we decided to use native PAGE. As mentioned in the previous section, TRPV3 forms a high molecular mass shoulder when eluting from a Superdex 200 column (Fig. 6A). We pooled and analyzed the TRPV3 shoulder and peak separately into a high and low molecular mass pool and analyzed them using native PAGE. Although dimers (denoted by an asterisk at ~ 200 kDa in Fig. 6) and tetramers (denoted by an arrow at ~ 400 kDa in Fig. 6) are detected in both pools, we found that the high molecular mass pool has a higher fraction of tetrameric assemblies and that the low molecular mass pool has a higher fraction of dimeric assemblies. No monomeric or trimeric species were detected. In addition, when the low molecular mass pool is collected conservatively avoiding any high molecular mass material (Fig. 6A, shaded area), and reanalyzed by gel filtration, the tetrameric shoulder is again observed (Fig. 6A, inset). We conclude that purified TRPV3 forms both dimers and tetramers in FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

Fig. 5. TRPV3 and TRPV3–GFP are monodisperse and can be purified to homogeneity. Following IMAC and Superose 6 gel filtration, proteins were purified on a Superdex 200 gel-filtration column. Elution profiles of TRPV3–GFP were collected by monitoring (A) absorbance at 280 nm and (B) fluorescence at an emission at 512 nm by excitation at 488 nm. (C) The elution profile of TRPV3 was collected by monitoring absorbance at 280 nm only. SDS/PAGE was used to assess the purity of TRPV3–GFP (A, inset) and TRPV3 (C, inset).

6015

S. Kol et al.


A

B

C

Fig. 6. Recombinant TRPV3 forms multimers and adopts an a-helical conformation. (A) TRPV3 elutes from a Superdex 200 pg gel-filtration column as a sharp, single peak, bearing a minor high molecular mass shoulder denoted by an arrow. The main peak containing the low molecular mass fraction of TRPV3 (shaded area) was reanalyzed by gel filtration (inset). (B) TRPV3 was divided into a high and low molecular mass pools and analyzed by native PAGE. Dimers and tetramers are denoted by an asterisk and an arrow, respectively. (C) Purified TRPV3 in FC12 detergent was analyzed by CD yielding a spectrum indicative of a folded protein of high a-helical content.

solution and that these species are in equilibrium. CD spectroscopy was used to investigate the overall fold of the purified TRPV3 channel. The far ultraviolet CD spectrum had negative peaks near 210 and 222 nm, consistent with a protein of high a-helical content (Fig. 6C), as would be expected for a protein with six transmembrane helices and additional a helices in the soluble domains. To examine whether purified TRPV3 has channel activity, we reconstituted it into planar lipid bilayers and measured representative current fluctuations across this bilayer. Although the bilayer exhibited no channel activity before addition of the protein, distinct single-channel fluctuations could be measured over the entire window of voltages immediately after addition of the TRPV3 proteoliposomes to the bilayer (Fig. 7A). From these measurements, we generated a single channel IV curve that exhibited however no significant rectification (Fig. 7B). To test whether these channel fluctuations are indeed generated by TRPV3 6016

we recorded channel activity at 32 °C before and after addition of 2-APB, a chemical agonist of TRPV3 activity [10,26]. Typical channel fluctuations were obtained after addition of TRPV3 proteoliposomes (Fig. 7C) with the unitary current amplitude corresponding to those obtained earlier. Addition to the bath medium of a saturating concentration of 2-ABP (400 lM), results after a delay of < 2 min in a strong activation of channel activity (Fig. 7D); this burst of channel activity was maintained throughout the duration of experiments (> 0.5 h). The burst-like activation of current fluctuations is most likely generated by the TRPV3 channel because a stimulating effect of 2-APB on the bilayer conductance was only observed after addition of TRPV3 proteoliposomes (data not shown). Collectively, the data suggest that the channel fluctuations reflect the activity of TRPV3, because channels with the same features were only observed in experiments with TRPV3 proteoliposomes and not with liposomes containing other channel proteins purified from E. coli (data not shown). As expected for TRPV3, the channel is Ca2+ permeable and has a low unitary conductance [5]. The general high open probability of the channel, which we observe here, may be explained by the absence of phosphatidylinositol (4,5) bisphosphate [27] in the membranes. Although consistent with an active TRPV3 channel, more data will be warranted to characterize and further confirm the functional integrity of the E. coli expressed human TRPV3 channel.

Discussion TRP channels have proven to be challenging structural biology targets. At present, the lack of a 3D structure prevents rational studies of the molecular mechanisms of TRP function, such as gating, modulation, ion permeation and selectivity. We describe a method to produce large amounts of active starting material for biochemical and X-ray crystallographic studies of a human TRP ion channel. We employed a highthroughput heterologous expression screen of human TRP channels in E. coli and identified TRPV3 as a favorable target for further optimization. After identification of a suitable detergent for extraction and purification, and upon optimization of buffer conditions, we analyzed the functional features of TRPV3 by CD spectroscopy, native PAGE and electrophysiology. Knowledge of how membrane proteins are synthesized in the cell is still very poor. For example, although each host cell appears to have a number of unique accessory factors required for membrane protein biogenesis, their precise roles are unclear. It is therefore unsurprising that the best production host FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

S. Kol et al.


A

Fig. 7. Proteoliposomes containing recombinant TRPV3 exhibit ion channel activity. (A) Representative current traces show the activity of the putative TRPV3 channel in DPhPC membranes at room temperature. Currents were measured in symmetrical solution with 100 mM CaCl2 10 mM Hepes–KOH pH 7.0. (B) The single channel IV curve exhibits no significant rectification. (C) Current trace recorded at +80 mV (temperature 32 °C) after addition of proteoliposomes with TRPV3; individual current fluctuations are shown at higher magnification on the right. (D) Addition of the chemical agonist 2-APB at 400 lM results in the same bilayer in a sustained increase in current fluctuations. The agonist was added ~ 80 s prior to the activity burst in (D).

C

D

frequently is the one most closely related in evolution to the source of the target membrane protein. For a membrane protein to be useful for subsequent studies in detergent-solubilized form, it should remain active or at least structurally well-defined in the detergent in which it is purified. Fos-choline detergents have frequently been used for the purification of TRP channels. For instance, FC12 was used to purify TRPV1 [21] and TRPA1 [13], and both were shown to be active in ligand-binding experiments and ligandinduced calcium efflux assays. In agreement with these studies, we have found that after screening many different detergents, we also find the Fos-choline class of detergents to be the most suitable for the E. coli expressed TRPV3 described here. CD spectroscopy was used to demonstrate folding of the purified TRPV3 and to estimate the secondary structure composition of the extracted, purified channel. We also probed the quaternary structure of TRPV3 using native PAGE. TRPV3 elutes from a SEC column as a single peak bearing a high molecular mass shoulder, which are assumed to represent dimers and tetramers, respectively. Indeed, the assumed tetramers reform from the dimeric species, which suggests equilibrium between these two forms. In many studies addressing the oligomerization properties of TRP FEBS Journal 280 (2013) 6010–6021 ª 2013 FEBS

B

channels tetramers were found, but prominent dimeric assemblies have also been observed [28–30]. Indeed, cross-linking with dimethyladipimidate using total cell homogenates yielded almost exclusively dimers of TRPV1 [31]. Other native PAGE experiments using purified TRPV1 in FC10 revealed almost only tetramers [21]. A high molecular mass shoulder was also observed during the purification of TRPA1 using FC12 and subsequent exchange into amphipols and this was explained by the formation of an adduct between protein and detergent [13]. It should be noted, that estimates of membrane protein oligomerization based on SEC are not very accurate due to the effects of detergent micelles on the mobility. It remains to be seen whether the observed mixture of dimers and tetramers has any physiological relevance or if it represents a detergent-induced artifact. Interestingly, TRP channels also form heterotetramers to diversify functional properties [32,33]. Using black lipid membranes we find that TRPV3 behaves consistently with functional characteristics observed for TRPV3 expressed in human cells [5]. Despite large, or even fundamental, differences in expression and membrane insertion of membrane proteins in bacteria, the efficiency of heterologous overproduction is very high, resulting in a yield of 6017

S. Kol et al.


1.2–1.5 mg of purified protein per liter of E. coli culture. The ability to produce the TRPV3 channel heterologously in E. coli in large amounts facilitates further biophysical and biochemical studies, such as by electrophysiological studies and X-ray crystallography to reveal the atomic details of TRPV channel structure and function. However, hitherto, initial crystallization attempts using purified TRPV3 have remained fruitless.

lysozyme, 1.0% DDM and 50 UmL 1 Benzonase (SigmaAldrich, St. Louis, MO, USA) and two freeze/thaw cycles. The lysate was centrifuged at 2500 g at 4 °C for 30 min to pellet cell debris and insoluble material. Recombinant proteins were purified using Ni2+ Chelating Sepharose FF (GE Healthcare, Piscataway, NJ, USA) in 1.2 lm filter plates as described elsewhere. Bound proteins were washed in lysis buffer containing 25 mM imidazole and eluted in lysis buffer containing 500 mM imidazole.

Experimental Procedures

Protein expression and membrane extraction

Cloning All enzymes were from New England Biolabs (Ipswich, UK). Primer design was aided by the Protein Crystallization Construct Designer at the Dutch Cancer Institute [34]. Primers were synthesized by TAG Copenhagen. All targets were named according to Uniprot and acquired either as cDNA clones [the Mammalian Gene Collection (MGC) and the ORFeome collection (Imagenes, Berlin, Germany)] or as synthetic genes (GeneArt, Regensburg, Germany). DNA sequences encoding the proteins of interest were PCR amplified with flanking LIC sites, T4 polymerasetreated and cloned into expression vector pNIC28-Bsa4 using published ligation-independent cloning protocols [35]. The plasmid pWaldo–GFPe–TRPV3 and pWaldo– GFPe–TRPV5 were created by adding XhoI 5′ and EcoRI 3′ restriction enzyme sites and subcloning into the appropriate sites in pWaldo–GFPe [18,19]. The endogenous EcoRI site in the TRPV3 gene was deleted using the 3′ primer. The nucleotide sequence of TRP cation channel subfamily V member 3 isoform 2 is available in the GenBank database under the accession number BC104868.

Small-scale protein expression and purification Expression strain colonies were inoculated in 96-well culture blocks containing TB supplemented with 50 lgmL 1 kanamycin and 25 lgmL 1 chloramphenicol. Cultures were grown overnight at 37 °C and 700 r.p.m. The next morning, 20 lL of the overnight cultures were diluted 50 times in TB supplemented with 50 lgmL 1 kanamycin and grown at 37 °C and 700 r.p.m. until the cell densities (D600) were 1.2–1.5. The plate was then moved to 18 °C and 700 r.p.m. for 30 min after which protein expression was induced by the addition of 0.5 mM isopropyl thio-b-Dgalactoside. Protein expression was then continued for 18 h after which cells were harvested by centrifugation. Cell pellets were disrupted by the addition of lysis buffer [100 mM Hepes–KOH, 500 mM NaCl, 10 mM imidazole, 10% glycerol, 0.5 mM tris(2-carboxyethyl)phosphine; pH 7.5] supplemented with Complete Mini EDTA-free protease inhibitor (Roche, Indianapolis, IN, USA), 1 mgmL 1

6018

The expression strain BL21 (DE3)-R3-pRARE2 was always transformed fresh. An overnight culture grown in TB medium supplemented with 50 lgmL 1 kanamycin and 25 lgmL 1 chloramphenicol was diluted 100 times into 3 L baffled flasks containing 1000 mL TB supplemented with 50 lgmL 1 kanamycin. Cells were grown at 37 °C and 180 r.p.m. until the cell density (D600) reached 1.2. The temperature was lowered to 18 °C and the cells were induced with 1 mM isopropyl thio-b-D-galactoside after 1 h. After 6 h the cells were harvested at 4000 g at 4 °C for 15 min, resuspended in 50 mM Hepes–KOH pH 7.5, 500 mM NaCl, 10% glycerol (before buffer optimization) or 50 mM Hepes–KOH pH 7.5, 150 mM KCl, 10% sorbitol (after buffer optimization) supplemented with Complete Mini EDTA-free protease inhibitor. Cells were lysed by three passes through a homogenizer at 10 000–15 000 psi, any debris and unbroken cells were removed by centrifuging twice at 18 000 g at 4 °C for 15 min. The total membrane extract was collected from the supernatant by centrifuging at 158 000 g at 4 °C for 2 h, resuspended in IMAC buffer A (20 mM Hepes–KOH pH 7.5, 500 mM NaCl, 10 mM imidazole, 10% glycerol or 20 mM Hepes–KOH pH 7.5, 150 mM KCl, 10 mM imidazole, 10% sorbitol, both supplemented with Complete Mini EDTA-free protease inhibitor) and frozen in N2 (l) or processed further. Inner membranes were prepared by loading total membranes onto a step gradient containing four-step sucrose gradient that consisted of 0.4, 1.0, 1.0 and 0.4 mL of a 36, 45, 51 and 54% (w/v) sucrose solution in 20 mM Hepes– KOH pH 7.5, respectively. Inner membranes were separated through centrifugation at 250 000 g at 4 °C for 30 min, collected from the gradient, and diluted with 5 vol. of 20 mM Hepes–KOH pH 7.5, 500 mM NaCl. Purified inner membrane vesicles were recollected by centrifugation at 158 000 g at 4 °C for 1 h, resuspended in IMAC buffer A and in frozen in N2 (l).

Detergent and buffer screening Inner membranes (250 lg) containing TRPV3–GFP were diluted in NaCl/Pi and solubilized with 100 lL of the Hampton Research detergent additive screen HR2-406 (end concentration 59 CMC). GFP fluorescence was measured


S. Kol et al.

before and after addition of detergent at an emission at 512 nm by excitation at 488 nm in a Varioskan Flash plate reader. After 2 h rotating at 4 °C, samples were transferred to centrifuge tubes and unsolubilized material was pelleted by centrifuging at 40 000 g at 4 °C for 30 min (Beckmann rotor type 42.2 Ti). The GFP fluorescence in the supernatant was measured again and expressed as a percentage of the original values before addition of detergents. Detergent screening was performed by solubilizing inner membranes in IMAC buffer A containing the appropriate detergent from Anatrace (Santa Clara, CA, USA). DDM and N-octyl-b-D-glucoside, octaethylene glycol monododecyl ether, LDAO, FC11, FC12, LFC12, DDDMG and DDMAB were used at a concentration of 1%, whereas DDMG and FC10 were used at a concentration of 2% because of their high CMC. Buffer screening was performed by solubilizing total membranes in IMAC buffer A containing 1% FC12. All other buffers contained three times the CMC of the appropriate detergent. Insoluble material was removed by centrifugation at 50 000 g at 4 °C for 20 min, the supernatant was loaded onto a HiTrap Chelating column (GE Healthcare), washed with 6% IMAC buffer B (20 mM Hepes–KOH pH 7.5, 500 mM NaCl, 10 mM imidazole, 10% glycerol) and eluted with 100% IMAC buffer B. The main elution fractions were pooled and concentrated using an Amicon Ultra-15 concentration cell (100 kDa cutoff). Analysis was performed on a Superose 6 HR 10/30 column (detergent) or on a Superdex 200 5/150 GL (buffer). Buffer scouting was performed by screening different pH values (6.0, 7.0, 7.5, 8.0 and 9.0), salts (150–500 mM NaCl or KCl) and additives (10–20% glycerol, sucrose or sorbitol).

Large-scale protein purification Total membrane preparations were solubilized by stirring for 1 h at 4 °C in IMAC buffer A (20 mM Hepes–KOH pH 7.5, 150 mM KCl, 10 mM imidazole, 10% sorbitol, Complete Mini EDTA-free protease inhibitor) containing 1% FC12. Insoluble material was removed by centrifugation at 50 000 g at 4 °C for 20 min, the supernatant was loaded onto a HiTrap Chelating column (GE Healthcare), washed with 6% IMAC buffer B (20 mM Hepes–KOH pH 7.5, 150 mM KCl, 500 mM imidazole, 10% sorbitol, 4.5 mM FC12) and eluted with 100% IMAC buffer B. The main elution fractions were concentrated using an Amicon Ultra-15 concentration cell (100 kDa cutoff) and loaded onto a HiLoad 16/60 Superose 6 gel-filtration column. The main TRPV3 peak was concentrated again and loaded onto a HiLoad 16/60 Superdex 200 pg gel-filtration column. Both columns were equilibrated with the 20 mM Hepes– KOH pH 7.5, 150 mM KCl, 10% sorbitol, 4.5 mM FC12. The final elution fractions were pooled and concentrated using an Amicon Ultra-15 concentration cell (100 kDa cutoff) at 15-min intervals with mixing in between. Final



concentrations ranged from 10 to 13 mgmL 1. Protein samples were frozen in N2 (l) and stored at 80 °C.

Protein analysis Protein samples were analyzed by Novex Bis–Tris 4–12% SDS/PAGE gels (Invitrogen, Frederick, MD, USA) and stained using Instant Blue Coomassie stain (Expedeon, Cambridge, UK) according to manufacturer’s instructions. Native PAGE was performed using the Bis–Tris gel system (Invitrogen) and gels were stained using the Coomassie R250 method. Where appropriate, gels were submitted to semidry electroblotting and immunodetection with a Penta His HRP conjugate antibody (Qiagen, Valencia, CA, USA) or an a-TRPV3 antibody (ab32734; Abcam, Cambridge, UK), followed by a secondary HRP-conjugated goat anti(rabbit IgG) (Proteintech, Chicago, IL, USA). Western blots were developed using the Immobilon Western kit (Millipore, Billerica, MA, USA). Chemiluminescence and in gel GFP fluorescence detection was done using the Fujifilm LAS-3000 imaging system (Fujifilm, Valhalla, NY, USA). The identity of recombinant proteins was checked by MALDI–TOF MS. Coomassie Brilliant Blue-stained protein bands were cut from the SDS gel, washed and digested by trypsin according to the instructions for the ProteoExtract All-in-One Trypsin Digestion Kit (Millipore).

CPM assay The CPM thermostability assay was carried out essentially as described previously [22,23]. In short, 25 lg of purified membrane protein was added to 150 lL of buffer containing 20 mM Hepes–KOH pH 7.5, 100 mM NaCl, and detergents at three times their CMC in a 96-well black Nunc plate. CPM dye (Sigma-Aldrich) at 4 mgmL 1 in dimethylsulfoxide was diluted 100-fold into buffer containing 20 mM Tris/ HCl pH 7.5, 100 mM NaCl, 0.03% DDM, and warmed to room temperature and 3 lL was added to the reaction. Fluorescence emission was measured at 463 nm (excitation 387 nm) on a Varioskan Flash plate reader at 40 °C. Recordings were measured every 5 min for 3 h with 15 s shaking interval between each reading. The fraction of folded protein at each time point was calculated by the quotient of raw fluorescence measured at each time point divided by the maximal fluorescence measured for the detergent/protein series.

CD spectroscopy Purified proteins were diluted 100 times in Milli-Q containing 3.0 mM FC12 and centrifuged prior to CD measurements performed on a Jasco J-815 spectrometer (Jasco, Easton, PA, USA). The sample volume in the 1 mm microcuvette (Hellma, Muellheim, Germany) was 200 lL. The protein concentration was 1.5 lM. For each sample five spectra were acquired for a wavelength range between 250

6019


and 190 nm in continuous scanning mode (20 nmmin 1, bandwidth 1 nm). Spectra were averaged and a blank measurement (gel-filtration buffer diluted 100 times with Milli-Q containing 3.0 mM FC12) subtracted.

Black lipid membrane experiments Planar lipid bilayers were formed by the monolayer folding technique [36] over a hole (~ 100 lm in diameter) in a Teflon foil (WAGNER). The hole was pretreated with a 1% hexadecane solution (Merck, Darmstadt, Germany) in n-hexane (Carl ROTH, Karlsruhe, Germany). In the next step, a lipid solution of 1,2-diphytanoyl-sn-glycero-3-phosphocholine (15 mgmL 1, from Avanti Polar Lipids, Alabaster, AL, USA) in n-pentane (Merck) was dispensed onto the experimental solution (100 mM CaCl2, 10 mM Hepes–KOH pH 7.0). After evaporation of the detergent, the solutions in both chambers were raised to form a lipid bilayer. After formation of a lipid bilayer, a voltage protocol was applied for at least 5–10 min to exclude contaminations or unspecific channel activity of, for example, lipid pores [37]. For incorporation of ion channels into the lipid bilayer proteoliposomes were added into the trans compartment. After fusion of the proteoliposomes with the lipid bilayer different voltage protocols were applied to record the resulting currents. All planar lipid bilayer measurements were carried out at a room temperature of 26–28 °C, unless noted otherwise. Ag/AgCl electrodes were connected to a head-stage of a current amplifier (L/M-EPC 7; List-Medical, Darmstadt, Germany). Membrane potentials are referred to the cis compartment. Single channel currents were filtered at 1 kHz and digitized with a sampling interval of 280 ls (3.57 kHz) by an A/D-converter (LIH 1600; HEKA Electronik, Lambrecht, Germany). Current traces at different membrane voltages were recorded via Patchmaster (HEKA), analyzed with Fitmaster (HEKA) and a lab-built software (KIELPATCH, www.zbm.uni-kiel.de/aghansen/software.html).

Acknowledgements This work was supported by the Novo Nordisk Foundation (SK and MS) and the LOEWE Soft Control initiative (CB and GT). Work in the PN laboratory was supported by the Lundbeck Foundation.

References 1 Ramsey IS, Delling M & Clapham DE (2006) An introduction to TRP channels. Annu Rev Physiol 68, 619–647. 2 Latorre R, Zaelzer C & Brauchi S (2009) Structure– functional intimacies of transient receptor potential channels. Q Rev Biophys 42, 201–246.

6020

S. Kol et al.

3 Clapham DE (2003) TRP channels as cellular sensors. Nature 426, 517–524. 4 Moran MM, McAlexander MA, Biro T & Szallasi A (2011) Transient receptor potential channels as therapeutic targets. Nat Rev Drug Discov 10, 601–620. 5 Xu H, Ramsey IS, Kotecha SA, Moran MM, Chong JA, Lawson D, Ge P, Lilly J, Silos-Santiago I, Xie Y et al. (2002) TRPV3 is a calcium-permeable temperature-sensitive cation channel. Nature 418, 181–186. 6 Smith GD, Gunthorpe MJ, Kelsell RE, Hayes PD, Reilly P, Facer P, Wright JE, Jerman JC, Walhin JP, Ooi L et al. (2002) TRPV3 is a temperature-sensitive vanilloid receptor-like protein. Nature 418, 186–190. 7 Peier AM, Reeve AJ, Andersson DA, Moqrich A, Earley TJ, Hergarden AC, Story GM, Colley S, Hogenesch JB, McIntyre P et al. (2002) A heat-sensitive TRP channel expressed in keratinocytes. Science 296, 2046–2049. 8 Montell C, Birnbaumer L, Flockerzi V, Bindels RJ, Bruford EA, Caterina MJ, Clapham DE, Harteneck C, Heller S, Julius D et al. (2002) A unified nomenclature for the superfamily of TRP cation channels. Mol Cell 9, 229–231. 9 Moqrich A, Hwang SW, Earley TJ, Petrus MJ, Murray AN, Spencer KS, Andahazy M, Story GM & Patapoutian A (2005) Impaired thermosensation in mice lacking TRPV3, a heat and camphor sensor in the skin. Science 307, 1468–1472. 10 Chung MK, Lee H, Mizuno A, Suzuki M & Caterina MJ (2004) 2-Aminoethoxydiphenyl borate activates and sensitizes the heat-gated ion channel TRPV3. J Neurosci 24, 5177–5182. 11 Gaudet R (2008) TRP channels entering the structural era. J Physiol 586, 3565–3575. 12 Gaudet R (2009) Divide and conquer: high resolution structural information on TRP channel fragments. J Gen Physiol 133, 231–237. 13 Cvetkov TL, Huynh KW, Cohen MR & MoiseenkovaBell VY (2011) Molecular architecture and subunit organization of TRPA1 ion channel revealed by electron microscopy. J Biol Chem 286, 38168–38176. 14 Mio K, Ogura T, Kiyonaka S, Hiroaki Y, Tanimura Y, Fujiyoshi Y, Mori Y & Sato C (2007) The TRPC3 channel has a large internal chamber surrounded by signal sensing antennas. J Mol Biol 367, 373–383. 15 Maruyama Y, Ogura T, Mio K, Kiyonaka S, Kato K, Mori Y & Sato C (2007) Three-dimensional reconstruction using transmission electron microscopy reveals a swollen, bell-shaped structure of transient receptor potential melastatin type 2 cation channel. J Biol Chem 282, 36961–36970. 16 Moiseenkova-Bell VY, Stanciu LA, Serysheva II, Tobe BJ & Wensel TG (2008) Structure of TRPV1 channel revealed by electron cryomicroscopy. Proc Natl Acad Sci USA 105, 7451–7455.


S. Kol et al.

17 Bill RM, Henderson PJ, Iwata S, Kunji ER, Michel H, Neutze R, Newstead S, Poolman B, Tate CG & Vogel H (2011) Overcoming barriers to membrane protein structure determination. Nat Biotechnol 29, 335–340. 18 Waldo GS, Standish BM, Berendzen J & Terwilliger TC (1999) Rapid protein-folding assay using green fluorescent protein. Nat Biotechnol 17, 691–695. 19 Drew DE, von Heijne G, Nordlund P & de Gier JW (2001) Green fluorescent protein as an indicator to monitor membrane protein overexpression in Escherichia coli. FEBS Lett 507, 220–224. 20 Phelps CB & Gaudet R (2007) The role of the N-terminus and transmembrane domain of TRPM8 in channel localization and tetramerization. J Biol Chem 282, 36474–36480. 21 Korepanova A, Pereda-Lopez A, Solomon LR, Walter KA, Lake MR, Bianchi BR, McDonald HA, Neelands TR, Shen J, Matayoshi ED et al. (2009) Expression and purification of human TRPV1 in baculovirus-infected insect cells for structural studies. Protein Expr Purif 65, 38–50. 22 Alexandrov AI, Mileni M, Chien EY, Hanson MA & Stevens RC (2008) Microscale fluorescent thermal stability assay for membrane proteins. Structure 16, 351–359. 23 Sonoda Y, Newstead S, Hu NJ, Alguel Y, Nji E, Beis K, Yashiro S, Lee C, Leung J, Cameron AD et al. (2011) Benchmarking membrane protein detergent stability for improving throughput of high-resolution X-ray structures. Structure 19, 17–25. 24 Eilers M, Patel AB, Liu W & Smith SO (2002) Comparison of helix interactions in membrane and soluble alpha-bundle proteins. Biophys J 82, 2720–2736. 25 Montell C (2005) The TRP superfamily of cation channels. Sci STKE 2005, re3. 26 Hu HZ, Gu Q, Wang C, Colton CK, Tang J, Kinoshita-Kawada M, Lee LY, Wood JD & Zhu MX (2004) 2-Aminoethoxydiphenyl borate is a common activator of TRPV1, TRPV2, and TRPV3. J Biol Chem 279, 35741–35748. 27 Doerner JF, Hatt H & Ramsey IS (2011) Voltage- and temperature-dependent activation of TRPV3 channels is potentiated by receptor-mediated PI(4,5)P2 hydrolysis. J Gen Physiol 137, 271–288.



28 Kedei N, Szabo T, Lile JD, Treanor JJ, Olah Z, Iadarola MJ & Blumberg PM (2001) Analysis of the native quaternary structure of vanilloid receptor 1. J Biol Chem 276, 28613–28619. 29 Hoenderop JG, Voets T, Hoefs S, Weidema F, Prenen J, Nilius B & Bindels RJ (2003) Homo- and heterotetrameric architecture of the epithelial Ca2+ channels TRPV5 and TRPV6. EMBO J 22, 776–785. 30 Garcia-Sanz N, Fernandez-Carvajal A, Morenilla-Palao C, Planells-Cases R, Fajardo-Sanchez E, FernandezBallester G & Ferrer-Montiel A (2004) Identification of a tetramerization domain in the C-terminus of the vanilloid receptor. J Neurosci 24, 5307–5314. 31 Jahnel R, Dreger M, Gillen C, Bender O, Kurreck J & Hucho F (2001) Biochemical characterization of the vanilloid receptor 1 expressed in a dorsal root ganglia derived cell line. Eur J Biochem 268, 5489–5496. 32 Schaefer M (2005) Homo- and heteromeric assembly of TRP channel subunits. Pflugers Arch 451, 35–42. 33 Hellwig N, Albrecht N, Harteneck C, Schultz G & Schaefer M (2005) Homo- and heteromeric assembly of TRPV channel subunits. J Cell Sci 118, 917–928. 34 Mooij WT, Mitsiki E & Perrakis A (2009) ProteinCCD: enabling the design of protein truncation constructs for expression and crystallization experiments. Nucleic Acids Res 37, W402–W405. 35 Gileadi O, Burgess-Brown NA, Colebrook SM, Berridge G, Savitsky P, Smee CE, Loppnau P, Johansson C, Salah E & Pantic NH (2008) High throughput production of recombinant human proteins for crystallography. Methods Mol Biol 426, 221–246. 36 Montal M & Mueller P (1972) Formation of bimolecular membranes from lipid monolayers and a study of their electrical properties. Proc Natl Acad Sci USA 69, 3561–3566. 37 Laub KR, Witschas K, Blicher A, Madsen SB, Luckhoff A & Heimburg T (2012) Comparing ion conductance recordings of synthetic lipid bilayers with cell membranes containing TRP channels. Biochim Biophys Acta 1818, 1123–1134.

6021