An uncharacterized region within the N-terminus of mouse TMC1 precludes trafficking to plasma membrane in a heterologous cell line

Mechanotransduction by hair cell stereocilia lies at the heart of sound detection in vertebrates. Considerable effort has been put forth to identify proteins that comprise the hair cell mechanotransduction apparatus. TMC1, a member of the transmembrane channel-like (TMC) family, was identified as a core protein of the mechanotransduction complex in hair cells. However, the inability of TMC1 to traffic through the endoplasmic reticulum in heterologous cellular systems has hindered efforts to characterize its function and fully identify its role in mechanotransduction. We developed a novel approach that allowed for the detection of uncharacterized protein regions, which preclude trafficking to the plasma membrane (PM) in heterologous cells. Tagging N-terminal fragments of TMC1 with Aquaporin 3 (AQP3) and GFP fusion reporter, which intrinsically label PM in HEK293 cells, indicated that residues at the edges of amino acid sequence 138–168 invoke intracellular localization and/or degradation. This signal is able to preclude surface localization of PM protein AQP3 in HEK293 cells. Substitutions of the residues by alanine or serine corroborated that the information determining the intracellular retention is present within amino acid sequence 138–168 of TMC1 N-terminus. This novel signal may preclude the proper trafficking of TMC1 to the PM in heterologous cells.

TMC1 and TMC2 are predicted to be membrane proteins 7 and localize at the tips of the shorter rows of stereocilia 16,17 , the site of MET channel activity 18 . However, TMC1 and TMC2 expressed in heterologous cell lines consistently fail to localize to the plasma membrane (PM) and instead are retained in the endoplasmic reticulum (ER) 6,7,19 . The failure of TMC1 to traffic to the PM in heterologous cell lines is a major obstacle to studying the structure and function of TMC1. PM localization might also be indispensable for the assembly of the multiprotein complexes of TMC1 and other essential MET channel proteins for structural and functional analyses.
Although the lack of a tissue-specific chaperone in heterologous systems is thought to be responsible for the intracellular retention of TMC1 7 , it is also plausible that the chaperone interaction region within TMC1 is left unbound in heterologous cells, thus being instead recognized as an ER retention and/or degradation signal by the subcellular machinery. Since most ER retention signals are usually located at the N-terminus or C-terminus of membrane proteins 20 , we reasoned the existence of such a signal could be prohibiting localization of TMC1 to the PM in heterologous cells. However, precise identification of these signals can be an arduous task, requiring the use of radioactive labeling and technically complex methodologies accompanied by hard-to-interpret results 21,22 . Interestingly, even though TMC1 seems to contain known ER retention signals such as KK motifs within its N-and C-termini, their ablation by alanine substitutions does not improve trafficking 7 . Several canonical ER retention motifs such as KDEL or KKXX signals have been identified 20,23,24 , but many more may exist 21,25,26 , which bioinformatic analysis alone cannot predict 27 . Thus, their existence remains to be characterized and de novo ones could be present within TMC1 amino acid sequence.
Here, we report the development and application of a straightforward and robust approach that could shed light into the intractable nature of TMC1 trafficking problems. This novel method allows for identification of uncharacterized intracellular retention signals within the N-and C-termini of a given membrane protein of interest. We took advantage of the ability of Aquaporin 3-GFP (AQP3-GFP) to display intense PM labeling in heterologous systems with very low cytoplasmic fluorescence [28][29][30] . We then used this construct as a PM localization reporter to pinpoint possible novel intracellular retention signals located at TMC1 N-and C-termini that completely abrogate any detectable PM localization of AQP3-GFP. When we tagged the 183 amino acid N-terminus of TMC1 with AQP3-GFP, this construct displayed an ER-like localization pattern and absence of any detectable PM labeling. We identified that the N-terminal region between amino acids 138-168 (TMC1 138-168 ) is responsible for intracellular retention and/or degradation. When tagged with AQP3-GFP, TMC1 138-168 prevents any PM labeling; while the resulting pattern is reminiscent of TMC1 localization in heterologous cells. In addition, TMC1 138-168 leads to a significant decrease in AQP3-GFP reporter intensity, suggesting that it may function as a degron. Substitution of TMC1 138-168 with a scrambled peptide, containing the same amino acids but in random order, abolishes the ability of this fragment to preclude trafficking to the PM. Moreover, substitutions of amino acids at the edges of TMC1 138-168 also resulted in robust PM labeling. Thus, TMC1 138-168 does indeed contain a sequence-specific ability, which precludes trafficking to the PM.

AQP3-GFP fusion protein displays unequivocally consistent membrane labeling in HEK293
cells. TMC1 is unable to properly traffic to the PM in heterologous systems when compared to other PM proteins such as Na/K ATPase (Fig. 1A). TMC1 gets invariably trapped at the ER in heterologous systems and never reaches the plasma membrane ( Fig. 1A right panels). To identify the TMC1 regions which are responsible for precluding trafficking to the PM, we searched for a membrane protein with an available luminal C-terminus that could demonstrate strong PM labeling when fused with GFP. We first assessed a palmitoylated version of GFP, which has the ability to be anchored to the cell membrane in order to achieve PM labeling, as previously reported 31 . However, the cell cytoplasm is also stained intensely, impeding any clarity in the localization pattern (Fig. 1B). In addition, we considered the usage of a farnesylated version of EGFP, EGFP-F 32 , but farnesyltransferases require an unoccupied C-terminus 33 , thus are not suited for this study. We then opted for Aquaporin-3 (AQP3), an innocuous obliged tetrameric polytopic membrane protein able to transport water as well as glycerol and other solutes across cell membranes 34 . AQP3-GFP fusion produces an intensely bright PM labeling in HEK293 cells with minimal residual cytoplasmic fluorescence (Fig. 1C,D), when compared to labeling using other membrane proteins such as the Na/K ATPase (Fig. 1A) or CD8 35 . Importantly, the C-terminus of GFP is left free at the luminal side. We used this AQP3-GFP construct as a reporter to investigate the inability of TMC1 to traffic to the PM in heterologous cells.
The 183 amino acid N-terminus of mouse TMC1 precludes AQP3-GFP trafficking to the plasma membrane. We first tested whether heterologously expressed full murine TMC1, fused to the C-terminal of AQP3-GFP, could successfully traffic to the PM ( Fig. 2A). We substituted residues of potential ER canonical retention signals of TMC1 with alanines in this and all subsequent constructs we tested. The AQP3-GFP-TMC1 displayed an ER-like localization pattern when expressed in HEK293 cells, similar to TMC1-YFP labeling (Figs 2B and 1A respectively). Thus, replicating the reticular pattern of TMC1, AQP3-GFP construct is suitable to identify regions responsible for halting trafficking to the PM.
To test whether the N-terminus of TMC1 (TMC1 1-183 ) contains such a region, TMC1 1-183 was fused to the C-terminus of AQP3-GFP. We estimated the length of TMC1 N-and C-termini based on the transmembrane prediction tool from the Department of Bio and Health informatics in the Center for Biological Sequence Analysis from the Technical University of Denmark. The generated prediction is in agreement with the most recently proposed topology of TMC1 8,9 . When expressed in HEK293 cells, AQP3-GFP-TMC1 1-183 labeling resembles the intracellular reticular pattern of AQP3-GFP-TMC1 and TMC1-YFP constructs (Figs 2C,B and 1A respectively). This suggests that, at least the N-terminus of TMC1, contains an uncharacterized region which prevent AQP3-GFP and presumably TMC1 from reaching the PM in heterologous systems. Next, we tested the C-terminus of TMC1 (TMC1 720-757 ). AQP3-GFP-TMC1 720-757 managed to reach the PM (Fig. 2D), suggesting the C-terminus of TMC1 does not preclude trafficking to the PM.
www.nature.com/scientificreports www.nature.com/scientificreports/ The N-terminal region of TMC1, encompassing amino acids 138-168, precludes trafficking to the plasma membrane. To narrow down the region within the N-terminus of TMC1 which halts trafficking to the PM, four constructs, each containing consecutive sequences of N-terminus, TMC1 1-46 , TMC1 47-92 , TMC1 93-137 , and TMC1 138-183 , were fused to the C-terminus of AQP3-GFP. When expressed in HEK293 cells, AQP3-GFP-TMC1 1-46 and AQP3-GFP-TMC1 47-92 constructs predominantly localized to the PM with some weak reticular pattern (Fig. 3A,B). This indicates that the first half of the N-terminus of TMC1 does not preclude trafficking to the PM. AQP3-GFP-TMC1 93-137 trafficked to the PM as well, although a distinct intracellular localization pattern was also noticed (Fig. 3C, asterisk). Next, AQP3-GFP-TMC1 138-183 was expressed in HEK293 cells. As a result, an unexpected drop in reporter intensity was observed, compared to the other three TMC1 N-terminal fragment containing constructs. This drop in expression was accompanied by a complete absence of PM localization (Fig. 3D). Thus, these results suggest that the fourth quarter of TMC1 N-terminus (amino acid residues 138-183) precludes membrane protein AQP3 from trafficking to the PM in HEK293 cells. Moreover, the ability of this region to significantly reduce the intensity of AQP3-GFP or just GFP ( Fig. 3D-F), is suggestive of a "degron" signal presence within TMC1 138-183 . To test whether any putative intracellular retention signals were perturbed by splitting TMC1 N-terminus into four parts, we generated two larger constructs containing the split regions, TMC1 1-92 and TMC1 1-137 . These fragments, containing approximately half and three quarters of the N-terminus of TMC1 respectively, were tagged with AQP3-GFP: AQP3-GFP-TMC1 1-92 and AQP3-GFP-TMC1 1-137 . When expressed in HEK293 cells, TMC1 1-92 and TMC1 1-137 did not prevent the AQP3-GFP from reaching the PM, but did display significant reticular localization patterns as well (Fig. 3G,H respectively).
Further, we aimed to narrow down the localization of the region residing within TMC1 138-183 that prevents trafficking to the PM. Hence, we tested three consecutive parts of the TMC1 138-183 region: TMC1 138-152 , TMC1 153-168 , and TMC1 169-183 . Interestingly, all of these three regions, when tagged with AQP3-GFP, reached the PM (Fig. 4A-C). We then tested two additional overlapping segments, TMC1 138-168 and TMC1 153-183 . When TMC1 138-168 was tagged with AQP3-GFP, it produced a significant reduction in reporter intensity (Fig. 4D, compared to those constructs described in Fig. 3D-F). Moreover, we did not observe any PM localization of AQP3-GFP-TMC1 138-168 , which is reminiscent of AQP3-GFP-TMC1 138-183 localization. In contrast, AQP3-GFP-TMC1 153-183 did reach the PM (Fig. 4E) without any decrease in reporter intensity levels. This suggests that, specifically, the region within amino acid residues 138-168 of TMC1 N-terminus precludes trafficking to the PM. TMC1 138-168 corresponds to AFKMMMAAAWAKFLRDFENFKAACVPWENKI (notice: italic AA substitutes potential ER canonical retention signal KK in the original sequence).
Since constructs TMC1 138-152 and TMC1 153-168 produced PM localization when tagged to AQP3-GFP reporter (Fig. 4A,B), we hypothesized that the region around the split, residues 152-153, is responsible for the intracellular retention. To test this, we substituted residues 149K, 150F, 152R, 153D, 154F, 155E, 156N, and 157F with alanine. However, such substitutions did not alter the intracellular retention activity of TMC1 138-168 (Fig. 5A). Further, we tested TMC1 142-168 and TMC1 138-163 (Fig. 5C,D). Neither shortened region precluded the reporter from reaching the PM, suggesting that amino acid residues at either end, but not in the middle of the TMC1 138-168 sequence, are critical in precluding trafficking to PM. Interestingly, a scrambled peptide containing the same amino acids as in TMC1 138-168 but in random order does not preclude AQP3-GFP from reaching the PM (Fig. 5B), suggesting that the activity of the TMC1 138-168 region is sequence-specific. www.nature.com/scientificreports www.nature.com/scientificreports/ TMC1 N-terminus with introduced mutations within TMC1 138-168 does not preclude trafficking to the plasma membrane. To investigate the ability of the TMC1 138-168 region to preclude trafficking to PM, we used alanine and serine mutagenesis to substitute both ends of the TMC1 138-168 sequence, AFKMM and WENKI amino acids, with ASASA and SASAS. When tagged with AQP3-GFP reporter, mutant TMC1 138-168 (mTMC1 138-168 ) reached the PM (Fig. 6A). We then introduced the same mutations within the entire N-terminus of TMC1. When tagged with AQP3-GFP, the mutant N-terminus of TMC1, containing the alanineand serine-substituted AFKMM and WENKI residues, AQP3-GFP-mTMC1 1-183 , was able to traffic to the PM of HEK293 cells (Fig. 6B). This is in stark contrast to wild type N-terminus, which precludes reporter trafficking to the PM (Fig. 2C).

Discussion
Members of the transmembrane channel-like family, TMC1 and TMC2, are key proteins in hair cell mechanotransduction, which enables hearing in vertebrates. TMCs are thought to form the permeation pathways of MET ion channel complexes 8,9 . Inability of TMCs to traffic to the PM in heterologous cell systems is a major obstacle to studying their structure and function. Although significant progress has been recently achieved, indicating TMC1 might form a dimer with predicted ion channel properties 8 , the trafficking issue is still a significant barrier impeding further work on the physiology and structure of the MET apparatus. AQP3-GFP-based reporter might provide a promising new tool to uncover protein regions, which preclude trafficking to the PM. When expressed in HEK293 cells, AQP3-GFP labels PM with unequivocal intense staining (Fig. 1C,D) and, importantly, with minimal residual cytoplasmic localization. When TMC1 is fused to the C-terminus of AQP3-GFP, the characteristic intracellular localization of TMC1 in HEK293 cells is not disrupted (Fig. 2B). This allowed us to use AQP3-GFP as a PM reporter in order to query TMC1 for regions, which preclude any detectable trafficking to the PM.  www.nature.com/scientificreports www.nature.com/scientificreports/ In this study, using HEK293 cells, we heterologously expressed the full and partial sequences of TMC1, tagged with AQP3-GFP. The heterologous expression revealed that the N-terminus of TMC1 and not the C-terminus, contains a region that prevents any detectable trafficking of AQP3-GFP reporter to the PM (Figs 2C,D, 3D-F and 4D). The sequence is positioned within amino acids 138-168, since expression of amino acids 138-152 or 153-168 of TMC1, when tagged with AQP3-GFP, reach the PM (Fig. 5). Moreover deletion of the five amino acids at either end of TMC1 138-168 , AFKMM or WENKI, abolish its ability to prevent trafficking to the PM (Fig. 5). Interestingly, when residues AFKMM and WENKI were substituted for ASAS and SASAS respectively, mTMC1 138-168 was able to reach the PM, as well as mTMC  (full N-terminus) (Fig. 6). TMC1 138-168 region also contains uncharacterized degron activity, which significantly decreases expression of AQP3-GFP, as well as soluble GFP (Fig. 3D-F). We further characterized TMC1 138-168 and found that it may contain a split signal composed of residues involving AFKMM and WENKI, both of which must be present to completely abrogate PM localization of AQP3-GFP reporter (Fig. 5C,D).
Nevertheless, TMC1 without 183 amino acid N-terminus still failed to reach the PM, when tagged with AQP3-GFP (data not shown). This result suggests that either lack of N-terminus results in a misfolded protein or additional ER retention motifs might still be present outside N-terminus. Since TMC1 has the ability to traffic to the PM at the tips of stereocilia, it is possible the ER retention signals that trap TMC1 in heterologous cells are masked by the action of specific protein partners in hair cells. www.nature.com/scientificreports www.nature.com/scientificreports/ Pure bioinformatical analysis of TMC1 might be unable to pinpoint the existence of uncharacterized ER motifs or any other regions responsible for halting trafficking to the PM 27 . Here we aimed to shed light into the possible signals, present within the N-and C-termini of TMC1, which may prevent TMC1 from trafficking to the PM in heterologous cell lines. Our data explains, at least partly, why TMC1 is retained intracellularly and fails to www.nature.com/scientificreports www.nature.com/scientificreports/ successfully traffic to the PM. AQP3-GFP could be further used as a PM reporter to examine putative transmembrane domains as well as inner and outer loops of TMC1 that could contain additional uncharacterized regions, which preclude trafficking to the PM. AQP3-GFP-based constructs could also be employed to study other membrane proteins that do not traffic to PM when expressed heterologously.
In conclusion, we have successfully applied AQP3-GFP-based Reporter (termed AGR) to effectively identify a hitherto uncharacterized region, which may contribute to the lack of PM localization of TMC1 in heterologous cell systems.

Materials and Methods
plasmid constructs. The sequence of turboGFP variant (Evrogen, Moscow, Russia) was fused to the C-terminal of human AQP3 (UniProtKB/Swiss-Prot #Q92482.2) in silico using Lasergene II software (DNASTAR, Madison, WI). The full cDNA containing AQP3-GFP was then synthesized by Genscript (Piscatawas, NJ). The cDNA was subcloned into a pcDNA5/FRT plasmid (Invitrogen, Carlsbad, CA) using restriction enzymes Hind3/ EcoRV. Each murine TMC1 (Pubmed #NP_083229.1) construct was also based on custom synthesis by Genscript with its known ER retention motifs KK or RRR substituted for alanines, and codon optimized for mammalian expression in HEK293 cells. Each Tmc1 based sequence was then subcloned into the AQP3-GFP-pcDNA5/FRT plasmid using Hpa1/EcoRV restriction sites. The palmitoylated GFP was custom synthesized, codon optimized for mammalian HEK293 expression and subcloned into pcDNA5/FRT plasmid using Nhe1/EcoRV restriction sites by Genscript. All constructs were confirmed by sequencing.
To express TMC1-YFP construct, the C-terminus of coding region of mouse Tmc1 was fused to the coding sequence of mVenus reporter using Hind III and BamHI sites in an mVenus-N1 vector 36 . Sequencing of this construct confirmed the identity as murine Tmc1 (Pumbed NM_028953).
Cell culture, transfection. HEK293 cells were purchased from Life Sciences (Carlsbad CA) and seeded in 50 mm glass-bottom dishes (WPI, Sarasota, FL) or 12-well glass-bottom culture plates (Cellvis, Mountain View, CA) and cultured in DMEM (Life Technologies, Carlsbad, CA) supplemented with 10% Fetal Bovine Serum and 1% pen/strep for 3 days prior to transfection. On the day of transfection, spent media was replaced by fresh media and plasmid constructs were transfected into HEK293 cells using Lipofectamine 2000 (Life Technologies, Carlsbad, CA) or K2 Transfection System (Biontex, Munchen, Germany) following the manufacturer's instructions. Red Fluorescent Protein based BacMam 2.0 constructs specific for ER were co-transfected the same day following manufacturer's instructions (Thermo Fisher Scientific, Waltham, MA). ER-Tracker ™ Red (Thermo Fisher Scientific) was used for live cell imaging as well (Fig. 1A only). To mark the plasma membrane in live cells, we used wheat germ agglutinin (WGA) CF594 conjugate (Biotium, Fremont, CA). To label the cell nucleus we used