General flexible nature of the cytosolic regions of fungal transient receptor potential (TRP) channels, revealed by expression screening using GFP-fusion techniques

Abstract

Transient receptor potential (TRP) channels are members of the voltage gated ion channel superfamily and display the unique characteristic of activation by diverse stimuli. We performed an expression analysis of fungal TRP channels, which possess relatively simple structures yet share the common functional characteristics with the other members, using a green fluorescent protein-based screening methodology. The analysis revealed that all the tested fungal TRP channels were severely digested in their cytosolic regions during expression, implying the common flexibility of this region, as observed in the recent structural analyses of the fungal member, TRPGz. These characteristics are likely to be important for their diverse functions.

Introduction

Transient receptor potential (TRP) channels are members of the voltage-gated ion channel superfamily, existing in animals and fungi.^1,2 They form cation channels with a generally non-selective pore at the center of the tetrameric assembly.³ TRP channels play various important physiological roles as environmental sensors and mediators of cellular signaling, by responding to wide arrays of extracellular and cytosolic stimuli, such as temperature, osmolarity change, chemical substances, cytosolic Ca²⁺ increase, and membrane potential.^1,3,4 Importantly, a particular TRP channel can often be activated by multiple stimuli. For example, TRPV1 is activated by heat >43°C, vanilloid compounds, and low pH,⁵ as well as membrane potential.⁶ In accord with these diverse functions, the TRP channel proteins are further categorized into seven subfamilies (TRPC, TRPV, TRPM, TRPA, TRPN, TRPP, and TRPML), which have distinct cytosolic structures, together with the conserved six transmembrane helices.^1,4 The cytosolic domains of TRP channels contain several functional regions responsible for channel regulation, such as the ankyrin repeat in TRPV1 for ATP- and calmodulin-mediated regulation,⁷ that in TRPA1 for activation by irritating chemical substances,^8,9 and the kinase domains in the TRPM proteins required for channel functions.¹⁰ Therefore, the cytosolic regions of the TRP channels are likely to be important platforms for channel regulation, underlying their diverse and multimodal responses.

Yeast and fungi possess an independent class of TRP channels from the aforementioned seven subfamilies, with a relatively shorter (and thus simpler) cytosolic region, when compared with their mammalian counterparts, and thus are regarded as being closer to the ancestral TRP channels.^1,4 A representative member from Saccharomyces cerevisiae, TRPY1, is a Ca²⁺-permeating vacuolar channel responding to extracellular osmotic upshock, and thus is considered to be important for protective reactions against environmental stimuli.^11,12 Recently, we identified a TRP channel from Gibberella zeae, TRPGz, and discovered that TRPGz is regulated by many diverse stimuli despite its simpler cytosolic architecture: the channel is activated by osmotic upshock, temperature change, oxidizer application, cytosolic Ca²⁺ increase, and membrane potential and is inhibited by phosphatidylinositol phosphate (PIPs).¹³ Notably, several of these functions are regulated by functional modules in the cytosolic region of TRPGz, such as the helix bundle module for activations by osmotic upshock and temperature increase, and the PIPs binding module for channel inhibition. These observations suggested that TRPGz, and probably other members of the fungal TRP channels as well, can serve as good model proteins to address the relationship between the cytosolic region and the diverse and multimodal regulations of channel gating.

In this study, we cloned several genes encoding fungal TRP channel homologs, and performed an expression screening using the green fluorescent protein (GFP)-fusion-based strategy. The GFP-fusion methodologies allow specific detection of the recombinant target proteins by GFP fluorescence without purification processes, and the strategies are recently often applied for sample evaluation and screening in the field of structural biology in combination with various biochemical methods.^14–20 In this study, we extended the GFP-fusion applications for a rapid determination of the proteolytic digestion sites in the proteins of interest, as well as the conventional application for the expression screening of the panels of fungus TRP channels. In combination with the expression and the amino-acid sequence analyses, we revealed the generally flexible nature of the cytosolic regions of fungal TRP channels, as observed in TRPGz.¹³ The results implied that the structural plasticity of the cytosolic regions underlies the various multimodal regulations of TRP channel functions.

Results

Identification and expression screening of fungal TRP channels

The PSI-BLAST search²¹ of the fungal TRP channels homologous to TRPY1 (YVC1, NP_014730) revealed the existence of several proteins with relatively higher sequence homology to TRPY1 (Table I), as reported.²² Some of the hit proteins had much longer amino acid sequences (more than 1000 residues), when compared with that of TRPY1. Therefore, we focused on the proteins with a similar number of amino acid residues to TRPY1 (675 amino acids), expecting that they would have relatively simpler architectures among the TRP channel family members. Multiple alignments of these eight proteins revealed that they have relatively high overall similarity to each other, with 108 identical amino acid residues (13.1%). The eight proteins shared similar transmembrane domains (∼20% identical sites), whereas their C-terminal cytosolic domains showed relatively low similarity (∼5% identical positions) (Fig. 1). We cloned the genes encoding these proteins, XP_448082 (Candida glabrata), XP_364983 (Magnaporthe oryzae), XP_504342 (Yarrowia lipolytica), and XP_660759 (Aspergillus nidulans) and performed their expression analysis together with four previously reported proteins, TRPY1, TRPY2, TRPY3, and TRPGz.^12,13,22

Table I.

The Top 15 Hits of the PSI-BLAST Search of Fungal Sequences, Using the TRPY1 Sequence^a

ID number	Source organism	Other name	Number of residues	Score (bits)	E-value	Identities (%)	Positives (%)	Gaps (%)	Ref.
NP_014730	S. cerevisiae	Yvc1p, TRPY1	675	–	–	–	–	–	12
XP_448082	C. glabrata	–	737	803	0.0b	62	77	0	c
XP_452815	K. lactis	TRPY2	676	767	0.0b	56	71	0	22
XP_716049	C. albicans	TRPY3, Ca019_2209	675	741	0.0b	42	62	4	22
XP_384354	G. zeae	TRPGz	692	735	0.0b	35	55	2	13
XP_754273	A. fumigatus	–	989	709	0.0b	34	54	9	–
XP_364983	M. oryzae	–	652	691	0.0b	36	58	4	c
XP_504342	Y. lipolytica	–	660	689	0.0b	40	60	5	c
XP_328311	N. crassa	–	1282	686	0.0b	30	51	11	–
XP_660759	A. nidulans	–	641	667	0.0b	33	51	9	c
XP_658170	A. nidulans	–	1096	634	3 × 10−180	17	34	17	–
XP_363612	M. oryzae	–	1348	573	8 × 10−162	16	33	15	–
XP_751014	A. fumigatus	–	956	567	5 × 10−160	17	35	13	–
XP_380532	G. zeae	–	1125	563	1 × 10−158	18	33	11	–
XP_389974	G. zeae	–	1315	551	3 × 10−155	16	35	15	–

^aThe result is after four PSI-BLAST iterations.

^bThe value was rounded down to 0 in the PSI-BLAST analysis.

^cThis study.

Figure 1.

Multiple sequence alignment and structural features of the fungal TRP channels analyzed in this study. Each protein is described in Table I. Putative transmembrane (TM) regions, TM1–TM6, and the helix bundle identified in TRPGz are indicated above the sequences. Boxed regions indicate the disordered regions predicted by DisEMBL.²³ The N-termini of the cleaved fragments are indicated by red triangles above the sequences [see Fig. 3 and Table II].

We expressed the aforementioned eight TRP channel homologs as C- or N-terminal GFPuv fusion proteins in Escherichia coli, and examined their expression profiles by in-gel fluorescence detection sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). This method allows the specific detection of the proteins of interest by the use of GFP-fluorescence, without any purification processes or immuno-detection.¹⁵ The in-gel fluorescence detection SDS-PAGE revealed that the expression of the full-length proteins was hardly detectable for all the tested homolog proteins, in both the membrane and cytosolic fractions (Fig. 2). The C-terminal fusions generated bands representing multiple small fragments, indicating severe proteolytic digestion [Fig. 2(A,B)]. On the other hand, the N-terminal GFP fusion constructs generated rather weak bands, implying that they were fragmented more severely or their expression levels were lower than those of the C-terminal GFP fusion products [Fig. 2(C,D)]. Notably, the disorder prediction of these proteins by DisEMBL²³ suggested that their N- or C-terminal cytosolic regions have very little secondary structure (Fig. 1). This observation is consistent with the SDS-PAGE results, which suggested that the proteolytic digestions of their cytosolic regions occurred in the E. coli cytosol or during the experimental process. All together, the results suggested that the cytosolic regions of the fungal TRP channels generally possess flexible structural characteristics.

Figure 2.

Expression profiles of fungal TRP channels fused with GFP at their C- (A, B) or N-termini (C, D) examined by in-gel fluorescence detection SDS-PAGE. The labels for the lanes are TRP channels from: (1) S. cerevisiae (TRPY1, NP_014730), (2) K. lactis (TRPY2, XP_452815), (3) C. albicans (TRPY3, XP_716049), (4) C. glabrata (XP_448082), (5) G. zeae (TRPGz, XP_384354), (6) M. oryzae (XP_364983), (7) Y. lipolytica (XP_504342), (8) A. nidulans (XP_660759), (M) markers. Panels (A, C) and (B, D) are the membrane- and soluble fractions of E. coli lysates expressing each construct, respectively. The red and green triangles represent the expected electrophoretic migration points for the full-length TRP channel homolog-GFPuv fusions and GFPuv alone, respectively. The results shown in panels (A, C) and (B, D) were from two different experimental trials among the multiple trials, in which the electrophoretic separation patterns for each trial were basically reproducible.

Determination of the proteolytic digestion sites in the cytosolic region of TRPGz

Similar to all the other tested proteins, the expression of full-length TRPGz also failed. Nevertheless, TRPGz fused with GFPuv at the C-terminal generated rather intense electrophoretic bands indicating the presence of several fragments in the molecular weight range of 23–65 kDa, which were larger than those from the other member proteins [Fig. 2(A,B)]. Interestingly, we found that intense fluorescence bands were observed in both the crude-membrane and soluble fractions, in the case of TRPGz [Fig. 2(A,B)].

We determined the digestion sites in the cytosolic region of TRPGz, by using the fact that the protein is the C-terminal GFP-fusion construct. As the estimated apparent molecular weights of the fragmented fluorescence bands suggested that the fragmentations had occurred within the C-terminal cytosolic region [Fig. 3(A,B)], we used the soluble fraction to identify the digestion sites. First, the cell lysate was concentrated by Ni-NTA metal affinity chromatography, and roughly fractionated by ion exchange chromatography [Fig. 3(A)]. The eluted or further electroblotted fragments were then identified by GFP fluorescence without any staining processes or immuno-detection [Fig. 3(C)], thus allowing direct N-terminal sequencing. The analyses revealed that the N-termini of these fragments were residues 539, 615, 649, 655, 672, and 673 (Table II and Fig. 1), indicating that the proteolytic digestions had occurred at these regions. These results were consistent with the flexible structural features of the C-terminal cytosolic region of TRPGz observed in our recent structural analyses; the region mostly consists of intrinsically disordered regions, with the multimerization region undergoing dynamic association and dissociation.¹³

Figure 3.

Identification of the digestion sites in the C-terminal cytosolic region of TRPGz. (A) In-gel fluorescence detection SDS-PAGE profile after NiNTA affinity purification of the C-terminal GFP fusion protein of TRPGz. (B) Anion exchange chromatography profile and (C) in-gel fluorescence detection SDS-PAGE of each fraction. Numbers (1)–(5) above the chromatography peak in panel B correspond to the lane numbers of the in-gel fluorescence detection SDS-PAGE in panel C. Fractions (1)–(4) were directly subjected to the N-terminal sequence analysis. Fraction (5), containing multiple bands, was further subjected to electro-blotting onto a PVDF membrane and detection by GFP fluorescence, and the arrow shows the band analyzed by N-terminal sequencing. The identified N-terminal sequences are shown in Table II.

Table II.

N-terminal Sequences of the Successfully Identified TRPGz Fragments

Fractions	N-terminal sequences	Residue number
1	T(669)SGDNKSGKK + S(668)TSGDNKSGK (G(671)DNKSGKKGK, D(672)NKSGKKGKK)a	669+668 (671, 672)a
2	A(655)DVLDTSDTP	655
3	S(649)ETTETADVL	649
4	S(615)QLGKTTAKG (L(614)SQLGKTTAK, G(618)KTTAKGSDD)b	615 (614, 618)b
5	A(539)WFEKRDARKI	539

^aSequences in parentheses represent minor sequences detected in the peak height of ∼10%.

^bSequences in parentheses represent minor sequences detected in the peak height of ∼20%.

Discussion

In this study, we performed the expression screening of fungal TRP channel homologs, by the use of GFP-based techniques. The GFP-based methodologies allowed quick parallel analyses of the expression profiles of the proteins, in terms not only of the expression levels but also of the expression length, such as partial expression or digestion during the recombinant expression.

The analyses demonstrated that the fungal TRP channels, with relatively simple cytosolic architectures, were highly susceptible to proteolytic digestion when expressed in E. coli (Fig. 2). Among the eight proteins tested, only TRPGz generated relatively strong electrophoretic bands that allowed us to identify the digestion sites by an N-terminal sequence analysis (Fig. 3). As the signals from the N-terminal GFP fusion constructs were relatively weak in the in-gel fluorescence detection SDS-PAGE [Fig. 2(C,D)], labile positions that are easily digested seem to exist in the N-terminal cytosolic region of all the tested channels, including TRPGz. This notion is also supported by the disordered region analysis by DisEMBL, which predicted disorder in the N-terminal portions of all the TRP channel homologs (Fig. 1). Another possible reason for the weak signals from the N-terminal fusion constructs is poor expression, due to improper folding and/or membrane targeting.

In the case of TRPGz, the digestion positions identified at 539, 615 (614, 618), 649, 655, 669, 668 (671, 672) are considered to be located at the borders of some structural or regulatory units. Many hydrophobic residues exist upstream of Residue 539 (Fig. 1), and thus it is considered to be the interface between the hydrophobic membrane-associating region and the hydrophilic cytosolic region. Interestingly, Position 615, where the most intense electrophoretic band was detected, is in the four-helix bundle structure, and its crystal structure was previously solved. Nevertheless, the cleavage at this position seems to be reasonable, because analytical ultracentrifugation and NMR measurements revealed that the helical structure exists in a dynamic equilibrium between monomer, dimer, and tetramer, with relatively weak interactions.¹³ The result supported the notion that the helix bundling region dissociates and associates even in the full-length channel. Interestingly, the heptad-repeat sequence found in the TRPGz helix bundle region seems to be conserved among the fungal TRP channels analyzed in this study (Fig. 1). Therefore, the helix-bundle or coiled-coil region for multimerization might be generally present in the C-terminal cytosolic regions of the fungal TRP channels. In addition, Residues 668 and 669 are very close to the basic residue cluster, which is involved in the inhibition of channel gating by TRPGz, by binding to acidic PIPs.¹³ This might be the reason why the C-terminal fragments of TRPGz were found in the membrane fraction of E. coli, as they may bind to the plasma membrane of E. coli containing acidic cardiolipins. Notably, no significant proteolytic digestion was observed downstream of Residue 669. This might be because the region was directly involved in binding to the lipid bilayer and thus was protected from digestion by proteases.

Previously, single particle analyses by electron microscopy have demonstrated that the cytosolic domains of the TRP channels possess unexpectedly large volumes for their molecular weights.^24–26 The TRP channels have multi-modular architectures within their cytosolic domains, and they are considered to work sometimes cooperatively, but at other times independently. We reported that TRPGz has modular architectures, such as the coiled-coil-like helix bundle, required for the osmotic shock response, and the PIPs binding domain, responsible for general channel inactivation, within the C-terminal cytosolic domain.¹³ Considering the present observations and the previously reported features of TRP channels, it is likely that the cytosolic domains of the fungal TRP channels behave mostly as intrinsically disordered proteins, rather than tightly folded globular proteins. These structural characteristics might be essential for the functions of the fungal TRP channels, which can be regulated by diverse stimuli to respond and induce protective cellular reactions on various cell stresses, as proposed for TRPY1 in S. cerevisiae.¹¹

During the review of our article, the first high-resolution structure of rat TRPV1, determined by electron cryo-microscopy, was reported.^27,28 We performed the disorder analysis of the rat TRPV1 by using DisEMBL, in a similar manner to the fungus TRP channels described earlier. Interestingly, we found that most of the cytosolic regions, except for the ankyrin repeat domain and the TRP box region, were predicted to be disordered: about 74% (170/230) of the amino acid residues belong to disordered regions (Fig. 4). These disordered regions were not included in the reported structure (PDB id: 3J5P), which had less than 5% of predicted disordered regions (23/587 residues). It is notable that the elucidated structure ends just before the predicted disordered region (residue number 720), although the protein constructs contained the downstream disordered region (Fig. 4). The analysis suggested that TRPV1 also possesses the flexible cytosolic regions with an intrinsically disordered nature, if we consider the entire protein, including the regions with structures that were not determined.

Figure 4.

DisEMBL²³ disorder analysis of rat TRPV1. The amino acid sequence is highlighted as follows: predicted disordered regions (red boxes); the region whose structure was solved by high-resolution cryo-electron microscopy (light blue), and the region included in the expression construct for the structural analysis (grey).²⁸ The structural elements observed in the electron microscopic structure, such as transmembrane (TM)-, TRP Box-, and ankyrin repeats-regions, are indicated above the sequence.

In addition to the expression profiles of the fungal TRP channels, we present a rapid determination method of the proteolytic digestion sites, applicable to the recombinant proteins assessed in the “GFP-based pipeline,”²⁹ in this study. The advantage of the GFP-fusion constructs is feasibility of the specific detection of even a small amount of the proteins of interest without any special detection method including immuno-detection, thus allowing the subjection to direct N-terminal analyses. Although here we performed the rough fractionation of the mixture of proteolytic products using ion exchange chromatography in advance of the N-terminal analyses (Fig. 3), we consider that the separation of the product mixture and the direct subjection to the analyses is basically practical by solely using electroblotting, enabling further quick determination.

Materials and Methods

Cloning and screening of fungal TRPY homolog genes

The genes encoding fungal TRPY homologs were identified by PSI-BLAST,²¹ using the TRPY1 sequence from S. cerevisiae. The genes encoding TRPY1 (NP_014730),¹² Y2 (from Klyveromyces lactis, XP_452815), and Y3 (from Candida albicans, XP_716049)²² were provided by Prof. Ching Kung, University of Wisconsin. The genes encoding the homologs from C. glabrata (XP_448082) and Y. lipolytica (XP_504342), without introns, were amplified by PCR from the genomic DNA provided by ATCC and Prof. T. Kanbe, Nagoya University, respectively. The genes encoding TRPY homologs from M. oryzae (XP_364983) and A. nidulans (XP_660759), as well as TRPGz from G. zeae (XP_384354),¹³ were amplified by PCR using the first-strand cDNA as templates, prepared from total RNA isolated from mycelia of each fungus.

All genes were inserted into the pcGFP_BC and pnGFP_BC vectors,¹⁸ provided by Prof. E. Gouaux, Vollum Institute, between the NcoI and NotI sites. E. coli C41(DE3) cells (OverExpress) bearing pLysS-RARE2 were transformed with the constructed expression vector for each TRP homolog. Five to 10 colonies were suspended in 3 mL LB medium, containing 1% glucose, 200 µg/mL carbenicillin, and 50 µg/mL chloramphenicol and cultured at 37°C for 2 h, and then IPTG was added at 1 mM to induce protein expression. After 4 h induction at 37°C, the cells were harvested by centrifugation, washed once with sonication buffer, containing 20 mM Tris-HCl, 200 mM NaCl, and 15 mM EDTA (pH 7.5), and then stored at −80°C until use.

Ice-cold sonication buffer (500 µL), supplemented with Protease inhibitor cocktail (Sigma P8849), was added to the frozen cell pellets, which were then subjected to 30 sonication pulses, using a Sonifier 250 (Branson) at Level 3, Duty 60, in an ice/CaCl₂ water bath. The sonicated samples were centrifuged at 21,500g for 30 min at 4°C. The obtained supernatants were further centrifuged at 42,000 rpm, using a TLA55 rotor (Beckman Coulter), for 30 min at 4°C. The supernatants were directly analyzed by SDS-PAGE. The membrane fractions were subjected to SDS-PAGE after solubilization with 20 mM n-dodecyl-β-d-maltoside (DDM). For solubilization, the membrane fractions were first suspended with a fire-polished Pasteur pipette and pestle in 200 µL of sonication buffer in a 1.5 mL tube, and then 50 µL of 100 mM DDM was added. The solution was gently rotated overnight at 4°C for complete solubilization.

SDS-PAGE was performed using Novex 4%–20% Tris-Glycine gels (Invitrogen), with samples treated with 5× SDS-PAGE sample buffer containing 2-mercaptoethanol. After electrophoresis, the gel was briefly washed with Elix water. The in-gel fluorescence derived from the fused GFPuv was detected with a ChemiDoc XRS system (Bio-Rad) equipped with a band pass filter for GFP detection (520 nm, FWHM 20 nm, Bio-Rad) and an LED epi-illuminating apparatus emitting 400 nm light, built in house.

Determination of the digestion sites in the TRPGz CTD

The C-terminal GFPuv-fusion protein of TRPGz was expressed using E. coli C41(DE3), cultured in NZCYM medium supplemented with 200 µg/L carbenicillin and 50 µg/L chloramphenicol. A 10 mL portion of the overnight culture of freshly transformed C41(DE3) cells was inoculated into 1 L of NZCYM medium, and protein expression was induced by 1 mM IPTG for 3 h. The E. coli cells from the 1 L culture were disrupted by three rounds of 50 sonication pulses with a Sonifier 250 (Branson) at Level 5, Duty 60, in an ice-water bath. After the debris was removed by centrifugation at 20,000g for 30 min at 4°C, the lysate was clarified by centrifugation at 190,000g for 1 h at 4°C. The cleared supernatant was then subjected to HisTrap column (GE Healthcare) chromatography. The bound proteins were eluted by 300 mM imidazole and were desalted by dialysis against 20 mM Tris-HCl (pH 7.5, 2 times for 1.5 h each, using 1000-fold buffer volume). The desalted protein was applied to a MonoQ anion exchange column (GE Healthcare, 1 mL column volume). During this purification procedure, the elution of the target protein was monitored by both the absorption at 280 nm and fluorescence at 507 nm excited by 395 nm, and peaks with fluorescent signals were specifically collected. The collected fluorescent fractions were examined by in-gel fluorescence detection SDS-PAGE. The fractions with a single fluorescent band were directly subjected to the N-terminal sequence analysis. The fractions with multiple fluorescent bands in the in-gel fluorescence were further subjected electro-blotting on a PVDF membrane, using an XCell SureLock Mini-Cell and XCell II Blot Module (Invitrogen) at room temperature, with transfer buffer composed of 12 mM Tris base and 96 mM glycine (pH adjustment and alcohol supplementation were not performed). After electro-blotting, the fluorescent band visualized by blue LED light was excised manually, and subjected to the N-terminal sequence analysis, performed using a Procise 494 protein sequencer (Applied Biosystems).