Elucidation of Molecular Impediments in the α6 Subunit for in Vitro Expression of Functional α6β4* Nicotinic Acetylcholine Receptors*

Background: Inefficient functional receptor expression in heterologous expression systems has hampered investigations of α6* nAChRs. Results: Determinants in the α6 subunit for α6β4* functionality have been delineated. Conclusion: Phe223 and the intracellular loop in α6 are molecular impediments to functional α6β4* nAChR expression in vitro. Significance: The molecular basis for the inefficient functional expression of α6β4* nAChRs in vitro has been elucidated. Explorations into the α6-containing nicotinic acetylcholine receptors (α6* nAChRs) as putative drug targets have been severely hampered by the inefficient functional expression of the receptors in heterologous expression systems. In this study, the molecular basis for the problem was investigated through the construction of chimeric α6/α3 and mutant α3 and α6 subunits and functional characterization of these co-expressed with β4 or β4β3 subunits in tsA201 cells in a fluorescence-based assay and in Xenopus oocytes using two-electrode voltage clamp electrophysiology. Substitution of a small C-terminal segment in the second intracellular loop or the Phe223 residue in transmembrane helix 1 of α6 with the corresponding α3 segment or residue was found to enhance α6β4 functionality in tsA201 cells significantly, in part due to increased cell surface expression of the receptors. The gain-of-function effects of these substitutions appeared to be additive since incorporation of both α3 elements into α6 resulted in assembly of α6β4* receptors exhibiting robust functional responses to acetylcholine. The pharmacological properties exhibited by α6β4β3 receptors comprising one of these novel α6/α3 chimeras in oocytes were found to be in good agreement with those from previous studies of α6* nAChRs formed from other surrogate α6 subunits or concatenated subunits and studies of other heteromeric nAChRs. In contrast, co-expression of this α6/α3 chimera with β2 or β2β3 subunits in oocytes did not result in efficient formation of functional receptors, indicating that the identified molecular elements in α6 could be specific impediments for the expression of functional α6β4* nAChRs.

The nicotinic acetylcholine (ACh) 2 receptors (nAChRs) mediate the rapid signaling of ACh and are widely distributed in the central nervous system (CNS) and in the periphery (1,2). The receptors are membrane-bound complexes assembled from five subunits, each consisting of a large extracellular N-terminal domain (NTD), a transmembrane domain (TMD) consisting of four transmembrane ␣-helices (TM1-TM4) connected by intracellular and extracellular loops, including a large second intracellular loop (ICL), and a short extracellular C terminus. Thus, the pentameric nAChR complex comprises three structural entities: an extracellular domain containing the orthosteric sites, a transmembrane domain containing the ion channel, and an intracellular domain, the three entities being assembled from the NTDs, the TMDs, and the ICLs of the five subunits, respectively (1,2).
The relatively promiscuous assembly of neuronal nAChRs from a total of eight ␣ (␣2-␣7, ␣9, and ␣10) and three ␤ (␤2-␤4) subunits gives rise to a plethora of physiologically relevant subtypes characterized by different distributions and distinct biophysical, kinetic, and pharmacological properties (1)(2)(3). The key roles played by this heterogeneous receptor population for cholinergic neurotransmission and for other neurotransmitter systems make them interesting as therapeutic targets in several neurodegenerative and psychiatric disorders (1,2,4).
The exploration of ␣6* nAChRs as putative drug targets has been hampered severely by the difficulties associated with efficient expression of functional receptors in heterologous expression systems (6,21). Several approaches have been applied to overcome this obstacle. First, co-expression of chimeric ␣6/␣3 or ␣6/␣4 subunits (␣6-NTD fused with ␣3or ␣4-TMD/ICL) with ␤2, ␤2␤3, and ␤4 subunits results in formation of functional receptors in both mammalian cells and oocytes (22)(23)(24)(25)(26)(27). Second, the minute responses observed for ␣6␤2␤3 and ␣6␤4␤3 nAChRs in oocytes have been found to be dramatically enhanced by the introduction of a ␤3 V273S mutant in the receptors (28,29). Finally, expression of functional ␣6␤2␤3 and ␣6␣4␤2␤3 nAChRs in oocytes has recently been accomplished by linking subunits in pentameric constructs; this concatemerization somehow makes up for the absence of whatever cellular factors that enables the formation of functional wild type (WT) receptors in neurons (26). Although these approaches have provided valuable tools for in vitro studies of ␣6* nAChRs, all of these are nevertheless modified receptors with the ever present uncertainty as to whether their functional characteristics diverge from those of WT ␣6* nAChRs (6).
In the present study, we further investigated the molecular determinants underlying the difficulties connected with in vitro expression of functional ␣6* nAChRs. A considerable number of novel ␣6/␣3 chimeras and several ␣6 and ␣3 mutants were constructed, and the functional properties of the receptors assembled from these subunits and various ␤ subunits in mammalian cells and Xenopus oocytes were characterized. Two molecular elements in the ␣6 subunit were identified as important determinants, or rather impediments, of the expression of functional ␣6␤4* nAChRs in heterologous expression systems.
Molecular Biology-The cDNAs of the ␣3, ␣6, ␤2, ␤3, and ␤4 nAChR subunits were amplified by the original vectors by PCR and subcloned into the pcDN〈3.1ϩ vector by use of the unique restriction sites NheI and XhoI (␣3, ␣6, ␤3, and ␤4) or NotI and XhoI (␤2). The chimeric ␣6/␣3 subunits were constructed using splicing by overlap extension PCR (30). This method was also used to insert a nucleotide sequence encoding for the c-myc epitope into ␣6, ␣3, and selected ␣6/␣3 chimeras and ␣6 mutants. The c-myc nucleotide sequence was inserted immediately downstream of the nucleotide sequence encoding for the signal peptide in each of the plasmids (␣6, -Val-Gly͉Cys 1 -Ala 2 -; ␣3, -Arg-Ala͉Ser 1 -Glu 2 -). Point mutations were introduced by using QuikChange site-directed mutagenesis according to the manufacturer's protocol (Stratagene, Santa Clara, CA). The integrity and the absence of unwanted mutations in all cDNAs created by PCR were verified by DNA sequencing (Eurofins MWG Operon, Martinsried, Germany).
Cell Culture and Transfections-The tsA201 cells were maintained in Dulbecco's modified Eagle's medium ϩ GlutaMAX TM -I supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C in a humidified 5% CO 2 atmosphere. The cells were split into 6-cm (1 ϫ 10 6 cells) or 10-cm (2 ϫ 10 6 cells) tissue culture plates and transfected the following day with a total of 4 g (6-cm plate) or 8 g (10-cm plate) of cDNA in a 1:1 ␣:␤4 ratio using PolyFect transfection reagent according to the protocol of the manufacturer (Qiagen, Hilden, Germany). The cells were used for the experiments 40 -48 h after the transfection.
Enzyme-linked Immunosorbent Assay (ELISA)-The ELISA was performed essentially as described previously (54). The tsA201 cells transfected with ␣3, myc-␣3, myc-␣6, myc-C1, myc-C2, myc-C6, or myc-␣6 F223L cDNAs together with ␤4 cDNA were seeded into poly-D-lysine-coated 24-well plates (3 ϫ 10 5 cells/well). The following day cells were washed in ice-cold wash buffer (phosphate-buffered saline (PBS) supplemented with 1 mM CaCl 2 ) and fixed in 4% paraformaldehyde (in PBS) on ice for 12 min. The following steps were performed at room temperature. The cells were washed three times with assay buffer and incubated with a blocking solution (3% dry milk in 50 mM Tris-HCl, 1 mM CaCl 2 , pH 7.5) for 20 min. After blocking, the cells were incubated with mouse anti-myc antibody (Invitrogen; diluted 1:500 in blocking solution) for 45 min. Then the cells were washed three times with wash buffer, incubated with blocking solution for 20 min, and incubated with goat anti-mouse horseradish peroxidase-conjugated (Invitrogen; diluted 1:400 in blocking solution) for 45 min. The cells were then washed three times in wash buffer before receptor expression was quantified using the 3,3Ј,5,5Ј-tetramethylbenzidine liquid substrate system (Sigma-Aldrich). The reaction was quenched with 1 N H 2 SO 4 after which the absorbance of the supernatants was determined at 450 nm. Total receptor expression levels of the respective myc-tagged subunits were determined by adding 0.1% Triton X-100 to the blocking solution used during the first round of blocking and the incubation with the primary antibody. Nonspecific binding was determined in parallel experiments on tsA201 cells expressing the WT (untagged) ␣3␤4 nAChR, and the "basal" staining determined in these wells was subtracted from the staining observed in the other wells.
Whole Cell Binding Assay-The whole cell [ 3 H]epibatidine binding experiments with tsA201 cells transiently expressing WT ␣3␤4, WT ␣6␤4, C1␤4, C6 F223L ␤4, and C16 F223L ␤4 nAChRs were performed essentially as described previously for whole cell [ 3 H]GR65630 binding to 5-HT 3 A receptors (31). The tsA201 cells were harvested in assay buffer (140 mM NaCl, 1.5 mM KCl, 2 mM CaCl 2 , 1 mM Mg 2 SO 4 , 25 mM HEPES, pH 7.4) using non-enzymatic cell dissociation solution (Sigma-Aldrich), counted, and divided into two equally sized fractions. Following centrifugation for 5 min, the resulting two cell pellets were resuspended to a concentration of 1 ϫ 10 7 cells/ml in assay buffer (intact cell population) or in assay buffer supplemented with 0.1% saponin (permeabilized cell population) and incubated for 5 min at room temperature. Visual inspection of the two cell populations mixed with trypan blue using a microscope confirmed that saponin treatment resulted in permeabilization of the cell membrane of virtually all cells (estimated Ͼ98%), whereas the cell membranes of virtually all non-treated cells were intact (estimated Ͼ98%). The samples were further diluted with assay buffer, and cells (1.5 ϫ 10 5 cells/reaction) were mixed with 3 nM [ 3 H]epibatidine in the absence (total binding) or presence of 300 M (S)-nicotine (nonspecific binding) in a total assay volume of 1 ml and incubated for 4 h at room temperature while shaking. Whatman GF/C filters were presoaked for 1 h in 0.2% polyethyleneimine, and binding was terminated by filtration through these filters using a 48-well cell harvester followed by washing with 3 ϫ 4 ml of ice-cold isotonic NaCl solution. Following this, the filters were dried, 3 ml of Opti-Fluor TM (Packard) was added, and the amount of bound radioactivity was determined in a scintillation counter. The binding experiments were performed in duplicate three to four times for each receptor.
FMP Assay-The FMP assay was performed in poly-D-lysinecoated, black 96-well plates (BD Biosciences). Transfected tsA201 cells were seeded into these plates 16 -24 h before the experiment. On the day of the experiment, the medium was aspirated, and the cells were washed with 100 l of Krebs buffer (140 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl 2 , 1.2 mM MgCl 2 , 11 mM HEPES, 10 mM D-glucose, pH 7.4). Then 100 l of Krebs buffer supplemented with FMP dye (0.5 mg/ml) was added to the wells after which the plate was incubated at 37°C in humidified 5% CO 2 for 30 min and assayed in a NOVOstar TM plate reader (BMG LABTECH, Offenburg, Germany) measuring emission at 560 nm (in fluorescence units) caused by excitation at 530 nm before and up to 1 min after addition of 33 l ACh solution in Krebs buffer. Experiments were performed in duplicate at least three times for each of the receptors. The concentration-response curves for ACh were constructed based on the differences in the fluorescence units between the maximal fluorescence levels recorded before and after addition of the agonist.
Electrophysiological Recordings-Electrophysiological recordings were performed using the two-electrode voltage clamp technique on Xenopus oocytes expressing the various receptors using a protocol adapted from previous studies (24,27,32). The oocytes were placed in a recording chamber continuously perfused with a saline solution (115 mM NaCl, 2.5 mM KCl, 10 mM HEPES, 1.8 mM CaCl 2 , 0.1 mM MgCl 2 ). Oocytes were clamped at Ϫ40 to Ϫ90 mV by a GeneClamp 500B amplifier (Axon Instruments, Union City, CA), and both voltage and current electrodes were filled with 3 M KCl. Six to eight different concentrations of the test compounds (in the saline solution described above) were applied until saturation followed by saline perfusion for 4 -6 min (C6 F223L ␤4, C1␤4, and WT ␣3␤4 recordings) or 2.5 min (C6 F223L ␤4␤3 recordings). Experiments were performed at room temperature on at least four oocytes from at least two different batches of oocytes for each subtype. Data were normalized to the maximum current elicited by ACh at the individual oocyte.
Data Analysis-All data analysis and curve fitting were performed using GraphPad Prism, version 5a (GraphPad Software, San Diego, CA). Concentration-response curves for agonists constructed based on the data obtained in the FMP assay and the oocyte recordings were fitted by non-linear regression using the equation for sigmoidal dose response with variable slope, where X represents the logarithm of the agonist concentration, Y represents the response, and "Top" and "Bottom" represent the plateaus in units of the y axis. Concentration-inhibition curves for mecamylamine in the oocyte recordings were fitted to a sigmoidal curve with variable slope using nonlinear regression, where X is the logarithm of the antagonist concentration, Y is the response, and Top and Bottom are the plateaus in units of the y axis. Specific binding in the [ 3 H]epibatidine whole cell binding experiments was defined as the difference between measured total and nonspecific binding. In the ELISA experiments, specific binding of anti-myc antibody was determined as the dif-ference between A 450 measured for the cells expressing the myc-tagged constructs and the A 450 measured for cells expressing WT (untagged) ␣3␤4 nAChR on the same plate.

Molecular Determinants in the ICL of ␣6 for the Expression of
Functional ␣6␤4 nAChRs-In a search for putative molecular elements in the ␣6 subunit underlying the problems obtaining efficient in vitro expression of functional ␣6* nAChRs, a series of 16 ␣6/␣3 chimeras (termed C1-C16) were constructed, coexpressed with the WT ␤4 nAChR subunit in tsA201 cells, and characterized functionally in the fluorescence-based FMP assay (Figs. 1-3 and Table 1).
In contrast to ␣6, the ␣3 subunit efficiently forms functional receptors in combination with ␤2, ␤2␤3, and ␤4 subunits in heterologous expression systems. Furthermore, it is the nAChR subunit most homologous to ␣6, making it ideal to use in this study. In concordance with the literature (22,24), ACh was found to elicit a robust functional response in tsA201 cells expressing the WT ␣3␤4 nAChR in the FMP assay, whereas no significant response could be detected in WT ␣6␤4-expressing cells (Tables 1 and 2 and Figs. 3 and 4). The WT ␣3␤4 and WT ␣6␤4 nAChRs were included as controls on all plates in the subsequent functional characterization of the receptors formed by chimeras C1-C16 in combination with WT ␤4. The dramatically different functionalities of the two WT receptors enabled us to relate the effects on ␣6␤4 signaling arising from various chimeric and mutant subunits to two fairly black-and-white references. The study was performed as an iterative process in which the results for chimeras obtained in one round formed the basis for the construction of additional chimeras to be studied in the next round.
The properties displayed by C1␤4 and C2␤4 strongly implicated the TMD and/or the ICL in ␣6 as domains containing "problem regions/residues" for assembly of functional ␣6␤4 nAChRs. To shed further light on these molecular elements, all chimeras subsequently generated comprised "pure" ␣6 NTDs and "mixed" ␣6/␣3 TMD/ICL regions (Fig. 1). Of the five chi-meras in the next round, only C3, C4, and C6 formed functional complexes with WT ␤4 (Table 1 and Fig. 3A). All of these chimeras contain an ICL composed completely of ␣3, and particularly informative was chimera C6 consisting of pure ␣6 NTD and TMD and a pure ␣3 ICL. In contrast, chimeras C5 and C7 with ICLs consisting completely of ␣6 did not form functional receptors with ␤4, further substantiating the notion of the ICL in ␣6 constituting a problem for the functional expression of ␣6␤4 nAChRs.
In the next round of chimeras, the ICLs of ␣6 and ␣3 were divided into three segments, a, b, and c, containing 21, 31, and 30 residues differing between ␣6 and ␣3, respectively (Fig. 2B). In the C8 -C13 chimeras, the a, b, and c segments from the two subunits were combined in various combinations, whereas the NTDs and TMDs of all chimeras were pure ␣6 ( Fig. 2A). Func- tional characterization of these chimeras co-expressed with WT ␤4 identified the c segment of the ICL as a particularly "problematic" segment for the expression of functional ␣6␤4 nAChRs, as the maximal responses elicited by ACh in cells expressing chimera C8 (␣6 a segment, ␣3 bc segments) and C9 (␣6 b segment, ␣3 ac segments) together with ␤4 were considerably higher than that for chimera C10 (␣6 c segment, ␣3 ab segments) ( Table 1 and Fig. 3B). The pattern of functionality was not completely black and white because C10␤4 was functional albeit very compromised compared with C6␤4 ( Fig. 3 and Table 1). On the other hand, the pattern observed for the C11␤4, C12␤4, and C13␤4 receptors supported a key role of the c segment for ␣6␤4 function. Here, C11 (␣6 ab segments, ␣3 c segment) was capable of forming functional receptors with ␤4, whereas cells expressing the C13␤4 (␣6 bc segments, ␣3 a segment) or C12␤4 (␣6 ac segments, ␣3 b segment) combinations were completely non-responsive to ACh ( Fig. 3C and Table 1).
In the final round of chimeras, the c segments were further subdivided into three segments, c1, c2, and c3, in a way so that each of the three segments contained 10 non-conserved residues between ␣6 and ␣3 (Fig. 2B). In the C14, C15, and C16 chimeras, the c1, c2, and c3 segments of ␣3 were introduced in ␣6, respectively ( Fig. 2A). Whereas ACh did not elicit agonist responses in cells transfected with the C14␤4 and C15␤4 combinations, a small but significant response was observed in C16␤4-expressing cells, identifying the c3 segment as an important region for functional expression of ␣6␤4 receptors (Table 1 and Fig. 3D). The 10 residues in the c3 segment of ␣6 not conserved in ␣3 were subsequently mutated to the respective corresponding ␣3 residues, and the mutants (␣6 D401E ,   ␣6 T419A , ␣6 V422I , and ␣6 E423Q ) were co-expressed with WT ␤4 in tsA201 cells and tested for functionality in the FMP assay. None of these mutant receptors exhibited a significant functional response to ACh exposure in the assay (data not shown). We did not attempt to further narrow down the molecular determinants for ␣6␤4 function in this segment. Molecular Determinants in TM1 of ␣6 for the Expression of Functional ␣6␤4 nAChRs-Although substitution of the ICL in ␣6 with that of ␣3 yielded functional receptors, the substantially smaller responses evoked by ACh through C6␤4 compared with C1␤4 indicated that TMD elements in ␣6 also could contribute to the poor in vitro functionality of ␣6␤4 nAChRs (Table 1). Although TM4 and the extracellular C terminus are the ␣6-TMD regions comprising most non-conserved residues compared with other ␣ nAChR subunits, the non-responsiveness of C7␤4 and the comparable responses evoked by ACh through C3␤4 and C1␤4 strongly suggested that any such elements are not harbored in these regions. Instead, the considerably smaller maximal response elicited by ACh through C4␤4 than through C1␤4 identified the six non-conserved residues in the TM1-TM3 region as candidates (Fig. 3A and Table 1). The non-responsiveness of C5␤4 containing ␣3 residues in four of these six positions as well as the findings in a recent study prompted us to focus on the two non-conserved residues in TM1: Leu 211 and Leu 223 in ␣3 corresponding to Met 211 and Phe 223 in ␣6, respectively. In this recent study, the maximal current amplitudes recorded from oocytes expressing ␣3 L211M ␤2 and ␣3 L223F ␤2 nAChRs were demonstrated to be significantly reduced compared with those of the WT ␣3␤2 nAChR (26). To investigate the importance of these two TM1 residues for ␣6␤4 nAChR function, the mutations L211M, L223F, and L211M/L223F were introduced in ␣3; the reverse M211L, F223L, and M211L/F223L mutations were introduced in ␣6, and the mutant subunits were co-expressed with WT ␤4 in tsA201 cells and characterized functionally in the FMP assay.
Analogously to the reported effect of the ␣3 L223F mutant on ␣3␤2 signaling (26), ␣3 L223F ␤4 displayed a significantly reduced maximal response compared with that of WT ␣3␤4 in the FMP assay. However, in contrast to the impaired signaling of ␣3 L211M ␤2 nAChR (26), introduction of the L211M mutation into ␣3 did not change the maximal response of ACh at the ␣3␤4 nAChR substantially ( Table 2). Co-expression of ␣3 L211M/L223F with ␤4 also resulted in the formation of receptors at which ACh exhibited a reduced maximal response compared with that at WT ␣3␤4, the R max value of the agonist at the double mutant being very similar to that at the ␣3 L223F ␤4 receptor (Table 2).
Strikingly, introduction of the F223L mutation in ␣6 resulted in the ability of the subunit to assemble into functional ␣6␤4 receptors ( Fig. 4A and Table 2). In contrast, the ␣6 M211L ␤4 combination did not display a significant functional response to ACh. Analogously to the pattern observed for the ␣3 mutants, the ␣6 M211L/F223L ␤4 receptor exhibited a functional response to ACh similar to that of ␣6 F223L ␤4.
Additive Effects of Molecular Determinants in ICL and TM1 in ␣6 for the Expression of Functional ␣6␤4 nAChRs-The observed rescue of ␣6␤4 nAChR function from introduction of even small ␣3 segments into the ICL as well as by a single mutation (F223L) in the TM1 of ␣6 prompted us to investigate whether the effects of these ICL and TM1 substitutions on ␣6␤4 function were additive. Introduction of the F223L mutation into the C6, C11, and C16 chimeras had dramatic augmenting effects on the functional properties of ACh at receptors containing all three chimeras, as the maximal responses exhibited by the agonist at C6 F223L ␤4-, C11 F223L ␤4-, and C16 F223L ␤4-expressing cells were more than double the size of those at C6␤4, C11␤4, and C16␤4, respectively (Table 3 and Fig. 4B).
Cell Surface Expression Levels of Chimeric ␣6/␣3 and Mutant ␣6 Subunits Co-expressed with WT ␤4 in tsA201 Cells-To elucidate to what extent the absolute number of receptors assembled in the cell membrane contributes to the respective func-  tionalities in tsA201 cells, the cell surface expression levels of selected receptors were determined. In the first line of experiments, myc-tagged versions of WT ␣6, WT ␣3, C1, C2, C6, and the ␣6 F223L mutant were co-expressed with WT ␤4 in tsA201 cells, and their expression patterns were investigated by ELISA. Insertion of the myc tag into ␣3 and C1 was found not to alter the functional properties of the ␣3␤4 and C1␤4 nAChRs (data not shown). Furthermore, the validity of the ELISA was verified in control experiments performed in parallel, where transfection of tsA201 cells with HA-tagged 5-HT3B was found not to result in significant cell surface expression, whereas co-expression of HA-5-HT3B with WT 5-HT3A gave rise to significant levels of cell surface expression of the HA-tagged subunit (54). As can be seen from Fig. 5A, tsA201 cells transfected with the myc-␣3␤4 and myc-C1␤4 combinations displayed significantly higher levels of "total" expression than cells expressing the myc-C2␤4 and myc-␣6␤4 receptors. Furthermore, myc-␣3␤4 and myc-C1␤4 displayed significantly higher cell surface expression than the myc-␣6␤4, myc-C2␤4, and myc-␣6 F223L ␤4 combinations, whereas the cell surface expression of myc-C6␤4 did not differ significantly (Fig. 5A). The relative cell surface expression, i.e. the percentage of the total number of myc-tagged subunits expressed at the cell surface, was very similar for five of the six receptors (40 -52%) with myc-C2␤4 being the outlier (18%). Interestingly, a distinct correlation was observed between the sizes of the maximal response evoked by ACh through the receptors in the FMP assay and their cell surface expression in the ELISA (Fig. 5A).
In another line of experiments, the number of binding sites for the orthosteric nAChR radioligand [ 3 H]epibatidine in tsA201 cells transfected with WT ␣3␤4, WT ␣6␤4, C1␤4, C6 F233L ␤4, and C16 F233L ␤4 nAChRs was determined in a whole cell binding assay using a saturating radioligand concentration (3 nM) and non-permeabilized and permeabilized cells (Fig. 5B). The number of [ 3 H]epibatidine binding sites at the surface of WT ␣6␤4-expressing cells was significantly lower than that for WT ␣3␤4-expressing cells, and all three receptors containing chimeric ␣6/␣3 subunits also displayed higher cell surface expression than WT ␣6␤4, albeit the C6 F233L ␤4 was the only receptor for which the difference was found to be significant (Fig. 5B).
The ELISA and whole cell binding experiments revealed a correlation between the cell surface expression levels of the receptors and their respective functionalities in the FMP assay. However, this correlation was not clear-cut, since some receptors with comparable levels of cell surface expression, for example C1␤4 and C16 F223L ␤4, displayed very different R max values in the functional assay (Tables 1 and 3 and Fig. 5). Furthermore, several receptors exhibiting a significant functional response to ACh in the FMP assay displayed surface expression levels similar to or only slightly higher than that of the non-functional WT ␣6␤4 (Fig. 5). Thus, although increased levels of cell surface expression of the receptors arising from the modifications introduced in the ␣6 subunit in some of these chimeras and mutants certainly seem to contribute to the functional rescue of WT ␣6␤4 function, the gain-of-function effects observed upon other ␣6 modifications cannot be ascribed to this factor. Functional Characterization of C6 F223L ␤4* and C6 F223L ␤2* nAChRs in Xenopus Oocytes-To investigate the functional properties of ␣6* nAChRs comprising one of the novel ␣6/␣3 chimeras in a more conventional assay for ligand-gated ion channels, C6 F223L , C1, WT ␣3, and WT ␣6 subunits were coexpressed with ␤4, ␤4␤3, ␤2, or ␤2␤3 subunits in Xenopus oocytes, and the assembled receptors were studied in two-electrode voltage clamp recordings.
Initially, we investigated whether the gain of function observed for C6 F223L ␤4 compared with WT ␣6␤4 in the FMP assay could be verified in the oocytes. Because of the extremely high expression levels of heterologously expressed proteins in this system, oocytes injected with WT ␣6␤4 cRNA actually form functional receptors, albeit agonist-evoked currents recorded from these have been reported to be minute (21,23,29). Thus, in contrast to the black-and-white functional rescue of ␣6␤4 function observed for the C6 F223L chimera in the FMP assay, a comparison of the functionalities of WT ␣6␤4 and C6 F223L ␤4 nAChRs in oocytes had to be based on the sizes of the maximal current amplitudes evoked by ACh in oocytes injected with comparable amounts of cRNA encoding for the two receptors. We observed a clear correlation between the amounts of WT ␣6␤4 cRNA injected into the oocytes and the current amplitude sizes evoked by 1 mM ACh in them. Upon injection of 70 -80 ng of cRNA of each subunit for WT ␣6␤4, maximal current amplitudes in the range of 300 -600 nA were observed upon application of 1 mM ACh (Table 4). In contrast, upon injection with 50 -60 ng of cRNA of each subunit, maximal current amplitudes of 20 -50 nA were recorded in two oocytes, whereas no currents could be measured in three other oocytes (Table 4). Because injection of similar amounts of C6 F223L ␤4 cRNA (50 -60 ng of cRNA of each subunit) in oocytes resulted in the formation of receptors responding robustly to ACh with maximal current amplitudes of up to 10 A, we conclude that the functionality of the ␣6␤4 nAChR in the oocyte expression system is also substantially augmented by the modifications introduced in the C6 F223L chimera.
Next we compared the ACh-evoked currents through the C6 F223L ␤4 nAChRs with those through WT ␣3␤4 and C1␤4 nAChRs. When similar amounts of cRNA for the WT ␣3␤4 and C6 F223L ␤4 combinations (20 -35 ng of each subunit) were injected into the oocytes, the maximal current amplitudes measured for C6 F223L ␤4 were consistently lower (50 -200 nA) than those recorded in oocytes expressing WT ␣3␤4 (up to 10 -15 M; Table 4). To obtain comparable maximal current amplitudes for all three receptor combinations in the following experiments, we injected double the amount of cRNA for C6 F223L ␤4 (50 -60 ng of each subunit) than for WT ␣3␤4 and C1␤4 (20 -35 ng of each subunit).
ACh elicited robust currents in a concentration-dependent manner in oocytes expressing the WT ␣3␤4, C1␤4, and C6 F223L ␤4 nAChRs (Fig. 6A). The ACh-evoked currents through C6 F223L ␤4 were efficiently eliminated by application of reference nAChR antagonists (ϩ)-tubocurarine (10 M) and mecamylamine (3 M) (data not shown). It should be mentioned that a pronounced decrease in maximal current amplitude was observed at ACh concentrations above 100 M in some of the C6 F223L ␤4-expressing oocytes, a phenomenon not observed for oocytes expressing WT ␣3␤4 and C1␤4 nAChRs (data not shown). The currents evoked by EC 20 ACh concentrations applied before and after the recording of currents for a range of different ACh concentrations differed somewhat in recordings at these oocytes. A decrease in current amplitude was observed for the EC 20 ACh application in the end of a run compared with that at the beginning, perhaps suggesting a more long lasting desensitization of this receptor than of WT ␣3␤4 and C1␤4. We nevertheless propose that the EC 50 value determined for ACh at C6 F223L ␤4 is a valid estimate of its actual potency at the receptor.
The physiological importance of the ␣6␤2␤3* nAChRs located on dopaminergic neurons in the midbrain prompted us to investigate the functional properties of these receptors expressed in oocytes. Although several different batches of cRNAs and oocytes were used in these experiments, applications of 1 mM ACh did not produce measurable responses in any of the WT ␣6␤2or C6 F223L ␤2-expressing oocytes tested (Table 4). In contrast, ACh elicited robust currents through WT ␣3␤2 with maximal current amplitudes in the 1-2-A range (Table 4). Interestingly, application of 1 mM ACh consistently produced significant currents in C6 F223L ␤2␤3-expressing oocytes, whereas the WT ␣6␤2␤3 nAChR was completely non-responsive to the agonist (Table 4 and Fig. 6B). The fact that measurable currents could be recorded at C6 F223L ␤2␤3 but not at C6 F223L ␤2 seems to be in concordance with previous reports of ␤3-mediated enhancement of ␣6* nAChR expression and function (23,33). However, the amplitudes of the currents recorded for C6 F223L ␤2␤3 were small (150 -250 nA) compared with those elicited by 1 mM ACh through the WT ␣3␤2␤3 and C1␤2␤3 nAChRs (Table 4).
Finally, we performed a detailed pharmacological characterization of the C6 F223L ␤4␤3 nAChR in oocytes (Fig. 6C). This subtype was chosen for these studies because the pronounced co-localization of ␣6 and ␤3 is suggestive of the presence of ␤3 in the majority of ␣6␤4* complexes in vivo (6). In these recordings, the duration of the intermediate saline perfusions between the drug applications was reduced from the 4 -6 min used in the C6 F223L ␤4 recordings to 2.5 min. We did not see the same degree of run-down in the C6 F223L ␤4␤3 recordings as for the C6 F223L ␤4 nAChR, which may also in part be ascribed to a stabilizing effect of ␤3 in the nAChR complex analogous to that observed previously for WT ␣6␤4␤3 and ␣6␤4 nAChRs (23). Six reference nAChR agonists were all found to evoke currents through the C6 F223L ␤4␤3 nAChR in a concentration-dependent manner (Fig. 6C). The rank order of agonist potencies at C6 F223L ␤4␤3 was (Ϯ)-epibatidine Ͼ sazetidine A Ͼ varenicline Ͼ (Ϫ)-cytisine ϳ (S)-nicotine Ͼ ACh. The current amplitudes evoked by sazetidine A through the receptor decreased dramatically at high concentrations (Ͼ3 M), a characteristic not observed for the other five agonists (data not shown). The maximal responses elicited by (S)-nicotine and sazetidine A through C6 F223L ␤4␤3 did not differ significantly from that evoked by ACh. In contrast, (Ϯ)-epibatidine was found to be a superagonist, and (Ϫ)-cytisine and varenicline displayed partial agonism at the receptor (Fig. 6C). Finally, ACh-evoked signaling through C6 F223L ␤4␤3 was antagonized in a concentrationdependent manner by the noncompetitive antagonist mecamylamine (Fig. 6D).

DISCUSSION
The inefficient expression of functional ␣6* nAChRs in heterologous expression systems has been the subject of extensive investigations addressing the origin of the problem and attempting to circumvent it by various approaches. In the present study, we have identified two molecular impediments in ␣6 for the functional expression of ␣6␤4* receptors: the Phe 223 residue in TM1 and the ICL (in particular the C-terminal part).
Because the focus of this study was on the ␣6 protein, it offers little insight into the putative neuronal factors or chaperones enabling expression of functional ␣6* receptors in vivo and does not address whether these are absent or compromised in vitro. Nevertheless, augmentation of ␣6␤4* function arising from ␣6 modifications has to be interpreted in light of the current understanding of nAChR trafficking and assembly. In an elegant study, a conserved PL(Y/F)(F/Y)XXN motif in the TM1s of the ␣1, ␤1, ␥, and ␦ subunits has been identified as a retention signal preventing the surface trafficking of unassembled subunits while being masked upon assembly into the muscle-type nAChR complex (43). Interestingly, the corresponding segment in ␣6 contains a methionine instead of highly conserved Leu residue (Fig. 7), and it has been speculated that this Met 211 residue could disrupt the retention signal in ␣6, thereby impairing the assembly of mature ␣6* receptors in the endoplasmic reticulum (26). However, although an Ala mutation of Leu 212 in the PLYFXXN sequence in ␣1 results in significantly decreased endoplasmic reticulum retention of the subunit (43), a Met residue in this position may not necessarily have a similar impact on endoplasmic reticulum retention, the Met residue being structurally more similar to Leu than Ala. Although the present study does not shed light on the role of Met 211 , introduction of the M211L mutation in ␣6 clearly does not rescue ␣6␤4 function, and thus the residue seems unlikely to be the sole molecular impediment for efficient functional expression of the receptors in vitro.
The Phe 223 residue located a couple of helix turns downstream of the TM1 retention signal is equally unique to ␣6 as Met 211 compared with other nAChR subunits (Fig. 7). The modest functionality of the ␣6 F223L ␤4 receptor could arise from an allosterically induced change in the conformation of the proximate retention motif or from a more direct effect of the introduced Leu residue on the assembly of the ␣6␤4 complex and/or its allosteric transitions. Based on the localization of the corresponding residues in high resolution structures of the Torpedo AChR and Cys-loop receptor orthologs (44 -46), Phe 223 is predicted to be positioned in the TMD subunit interface of the ␣6* complex facing toward TM3 of the neighboring subunit. The Cys-loop receptor TMD subunit interface is a hot spot for allosteric modulation (1), and a molecular change in this region could be speculated to result in a receptor that is more responsive to agonist stimulation.
The C-terminal part of the ICL in ␣6 is likely to present a different molecular hindrance to functional expression of ␣6␤4* receptors than Phe 223 . First of all, because of the sheer distance between the two molecular elements, it would be difficult to imagine modifications in this loop having an effect on the retention motif in TM1. Second, the contributions of deletions of the two elements to the enhancement of ␣6␤4 functionality appears to be additive (Table 3 and Fig. 4B). A role of the ␣6-ICL for the inefficient expression of functional ␣6␤4 receptors is not surprising considering reported involvement of ICLs in the trafficking, expression, and signaling of other nAChRs through their interactions with intracellular proteins (47)(48)(49)(50). However, the molecular impediments to functional expression of ␣6␤4 receptors comprised within the ICL are certainly less defined than Phe 223 , as we have not been able to pinpoint the problem to a specific residue or motif in the loop. Although substitution of the non-conserved residues contained in the C-terminal Asp 412 -Trp 437 segment of the ␣6-ICL with the corresponding ␣3 residues results in a functional receptor (C16␤4), the significantly higher maximal responses elicited by ACh through C11␤4 and C6␤4 and the small but significant response evoked through C10␤4 could indicate that the entire ICL constitutes a molecular obstacle to functional receptor expression. Alternatively, introduction of an ␣3 segment instead of a segment in the ␣6-ICL region that does not in itself constitute a problem could induce a conformational change in the C-terminal part of the loop and thereby diminish the impact of a specific problematic molecular element located here.
In agreement with a previous study of WT ␣6␤4 and ␣6 NTD / ␣4 TMD/ICL ␤4 nAChRs (24), the receptors formed by the surrogate ␣6 subunits C1, C6 F223L , and C16 F223L with ␤4 were found to exhibit higher cell surface expression levels than WT ␣6␤4 (Fig. 5B). However, although this definitely seems to be an important component of the augmented functionality of several of the receptors in this study, increased trafficking and/or incorporation of the subunits into receptor complexes in the cell membrane does not account for the gain-of-function effects arising from all ␣6 modifications. Thus, introduction of the Leu 223 residue and/or an ␣3-ICL segment in ␣6 may also alter the allosteric transitions of the receptor, induce another subunit stoichiometry in the complex, or in other ways affect its functionality. Whatever the molecular mechanisms causing the augmented functionality of the ␣6␤4* receptors containing these surrogate ␣6 subunits are, it is important to remember that neither Phe 223 nor the C-terminal ICL segment in ␣6 constitute an insurmountable hindrance for expression of functional receptors in neurons. Thus, these so-called molecular impediments in ␣6 are really only in vitro manifestations existing in light of the deficiency of the heterologous expression system to efficiently express functional WT ␣6␤4* nAChRs.
Interestingly, the C6 F223L chimera exhibits strikingly different efficiency when it comes to the formation of functional ␣6␤2* and ␣6␤4* receptors. Although the minute currents elicited by ACh through C6 F223L ␤2␤3 can be considered a gain-offunction effect compared with the completely non-responsive WT ␣6␤2␤3, the molecular modifications introduced in ␣6 to facilitate functional expression of ␣6␤4* receptors clearly do not translate into nearly as an efficient rescue of ␣6␤2* function. In this respect, C6 F223L differs from the classical ␣6 NTD / ␣3 TMD/ICL chimera (C1), but analogously the ␣6 NTD /␣4 TMD/ICL chimera has been shown to form functional receptors with ␤4 but not with ␤2 (24), and co-expression of this chimera with ␤2␤3 yields functional receptors (27). Furthermore, a complex pattern of subunit compatibilities has been observed for hybrid nAChRs formed from human and murine ␣6, ␤2, ␤4, ␤3, and ␤3 V273S subunits (29). All these findings bear witness to the allosteric nature of the nAChR complex and illustrate one of the potential shortcomings of the surrogate ␣6 subunit: although the Leu 223 residue and the ␣3-ICL in C6 F223L appear to have overcome the inborn molecular impediments in ␣6 for assembly and expression of functional ␣6␤4* nAChRs, other or additional elements in the subunit may counteract efficient formation of functional ␣6␤2* receptors.
In conclusion, it is important to stress that we do not consider the novel ␣6/␣3 chimeras presented in this study to be superior to other surrogate ␣6 subunits or other approaches used to express functional ␣6* nAChRs in vitro in previous studies. The higher ␣6 content in the C6 F223L and C16 F223L chimeras compared with the classical ␣6 NTD /␣3 TMD/ICL and ␣6 NTD /␣4 TMD/ICL chimeras may be considered an advantage for example when it comes to screenings for novel ␣6␤4* ligands. On the other hand, the inefficient formation of functional ␣6␤2* nAChRs from the chimeras clearly reduces the overall utility of the constructs. Furthermore, although the pharmacological properties exhibited by the C6 F223L ␤4␤3 nAChR seem to be in good agreement with previous findings for ␣6␤4* and other nAChRs, the characteristics of these receptors cannot be assumed to mimic those of WT ␣6␤4* nAChRs on all accounts, especially when considering the important role of the ICL in the Cys-loop receptor for its trafficking, assembly, and biophysical properties (51)(52)(53). Such concerns will inevitably exist for any ␣6* nAChR assembled from modified ␣6 subunits or concatamers, and thus the identification of the neuronal factors or chaperones enabling the expression of functional receptors in vivo and the resulting ability to express functional WT ␣6* nAChRs in heterologous expression systems would constitute a major leap forward in this field.