A human-specific, truncated α7 nicotinic receptor subunit assembles with full-length α7 and forms functional receptors with different stoichiometries

The cholinergic α7 nicotinic receptor gene, CHRNA7, encodes a subunit that forms the homopentameric α7 receptor, involved in learning and memory. In humans, exons 5–10 in CHRNA7 are duplicated and fused to the FAM7A genetic element, giving rise to the hybrid gene CHRFAM7A. Its product, dupα7, is a truncated subunit lacking part of the N-terminal extracellular ligand-binding domain and is associated with neurological disorders, including schizophrenia, and immunomodulation. We combined dupα7 expression on mammalian cells with patch clamp recordings to understand its functional role. Transfected cells expressed dupα7 protein, but they exhibited neither surface binding of the α7 antagonist α-bungarotoxin nor responses to acetylcholine (ACh) or to an allosteric agonist that binds to the conserved transmembrane region. To determine whether dupα7 assembles with α7, we generated receptors comprising α7 and dupα7 subunits, one of which was tagged with conductance substitutions that report subunit stoichiometry and monitored ACh-elicited channel openings in the presence of a positive allosteric α7 modulator. We found that α7 and dupα7 subunits co-assemble into functional heteromeric receptors, which require at least two α7 subunits for channel opening, and that dupα7's presence in the pentameric arrangement does not affect the duration of the potentiated events compared with that of α7. Using an α7 subunit mutant, we found that activation of (α7)2(dupα7)3 receptors occurs through ACh binding at the α7/α7 interfacial binding site. Our study contributes to the understanding of the modulation of α7 function by the human specific, duplicated subunit, associated with human disorders.

␣7 is a homomeric member of the nicotinic receptor (nAChR) 2 family, which belongs to the pentameric ligand-gated ion channel superfamily (1)(2)(3). ␣7 receptors are localized in the central and peripheral nervous systems as well as in nonneuronal cells. They have pleiotropic effects ranging from the modulation of neurotransmitter release and the induction of excitatory impulses in the nervous system to the regulation of inflammatory responses in the immune system (4,5). Decreased expression and function of ␣7 has been associated with neurological and neurodegenerative disorders, including Alzheimer's disease, schizophrenia, bipolar disorder, attention deficit hyperactivity disorder, and autism spectrum disorder (6).
Nicotinic receptors contain a large extracellular domain, which carries the agonist-binding site, a transmembrane region, which is formed by four transmembrane segments of each subunit (M1-M4) with the M2 domains forming the walls of the ion pore, and an intracellular region that contains sites for receptor modulation and determinants of channel conductance (3,5). At the interface between the extracellular and transmembrane domains, several loops form a network that relays structural changes from the binding site toward the pore. This region, named the coupling region, contributes to the fundamental mechanism of receptor activation (7)(8)(9).
The acetylcholine (ACh)-binding sites are located at the interfaces of the extracellular domains of adjacent subunits. Each binding site is composed of a principal face provided by one subunit, which contributes three loops, named A, B and C, and a complementary face provided by the adjacent subunit, which contributes loops D, E, and F (10,11). The homomeric ␣7 receptor has five identical ACh-binding sites; however, ACh occupancy of only one site is enough for activation (12).
The ␣7 receptor subunit gene, CHRNA7, has 10 exons and is located on the long arm of chromosome 15 (15q13-q14). A hybrid gene, CHRFAM7A, has arisen from a relatively recent partial duplication that comprises exons 5 to 10 of the CHRNA7 gene and is positioned in the same chromosome, centromeric to the CHRNA7 gene by 1.6 Mb, interrupting the genetic element FAM7A. Interestingly, the CHRFAM7A gene is human specific (6,13,14). The final protein, dup␣7, is a truncated receptor subunit that lacks the first 95 amino acid residues of the ␣7 protein, which includes loops D and A of the agonistbinding site, and instead contains 27 amino acid residues from FAM7A at the N-terminal domain (15). The CHRFAM7A gene is located in a complex region on chromosome 15 that includes many segmental, highly variable, duplications that result in several different copy number variants. These multiple polymorphisms are associated with the risk to develop several neurological and psychiatric disorders, such as schizophrenia, bipolar disorder, autism, and idiopathic epilepsies (16,17). Among the copy number variants present in chromosome 15, another variant-truncated subunit gene has been described, which shows a 2-bp deletion in exon 6 of the CHRFAM7A, which has been associated with schizophrenia and P50 sensory gating deficit (13,16).
High expression of dup␣7 also occurs on human leukocytes (17,18). In primary monocytes and macrophages, lipopolysaccharide treatment down-regulates the expression of dup␣7 (19). These findings suggest that the duplicated isoform might have a role in the immune system and cholinergic anti-inflammatory pathway. It is also present in a great variety of epithelial cells, confirming a wide distribution of expression of this truncated subunit (20).
The dup␣7 subunit has been expressed heterologously in mammalian cell lines and Xenopus oocytes in a few studies. Immunological studies in GH4C1 cells suggested that dup␣7 can reach the cell periphery, although it appeared to be at a more inner location than that of ␣7 (21). No macroscopic responses have been detected after exposure to ␣7 agonists (21,22). The role of dup␣7 remains unclear because reduction of ␣7 currents, compatible with a negative modulator role, has been shown in oocytes (21,23) but not in neuronal cells (22). By incorporating fluorescent tags Wang et al. (22) proposed that dup␣7 can assemble with the full-length ␣7 because both subunits are membrane localized in close position. The stoichiometry of these hybrid receptors and their function remain unknown.
To find a way to identify the possible pentameric ␣7/dup␣7 arrangements and determine their functional signature, we combined cell expression and single-channel recordings with the electrical fingerprinting strategy (12, 24 -26). Receptors were generated using combinations of ␣7 and dup␣7 subunits, one of which carries a reporter conductance mutation that allows defining the subunit stoichiometry of the receptor that originated each single-channel opening event in real time. Our results provide novel information about the function of the duplicated ␣7 subunit, which may have a significant role in immunomodulation and in the pathophysiology of neurological disorders.

Heterologous expression of dup␣7 on mammalian cells
The cloned dup␣7 cDNA includes the 1236-nucleotide sequence previously deposited in NCBI and corresponds to the CHRFAM7A isoform 1 (accession number NM_139320). The initiator methionine in exon B of CHRFAM7A will produce a protein including amino acid residues coded by part of exon B, exon A of the FAM7A gene, and exons 5-10 of CHRNA7 (Fig.  1A). The final protein contains the N-terminal domain with 27 residues of FAM7A followed by the ␣7 sequence starting at amino acid residue 96 and therefore lacks loop A and D of the ACh-binding site (Fig. 1A).
To establish that the cloned dup␣7 cDNA is translated into the truncated subunit in our system, BOSC-23 cells were transfected with plasmid vectors containing either dup␣7 or ␣7 cDNAs together with plasmids encoding the chaperone proteins, Ric-3 and NACHO in a 2:1:1 molar ratio, respectively. The expression of ␣7 and dup␣7 subunit proteins was detected by Western blotting from whole cell lysates using a previously characterized antibody raised against the intracellular M3-M4 loop that is common to both subunits (27). Western blotting revealed bands in the 55-57-kDa range in ␣7 expressing cells and in a lower molecular mass range (45)(46)(47)(48)(49)(50) in those transfected with dup␣7 cDNA. These bands, which were not detected in nontransfected cells, are compatible with the expected molecular weights of both subunits (Fig. 1B) (18,19). Double bands as the ones observed have been described before for ␣7 and might represent forms differing in glycosylation or in other post-translational modifications (28 -30). Thus, our Western blotting results confirm that dup␣7 is transcribed and translated in our heterologous expression system.

dup␣7 reduces the number of ␣-BTX-binding sites when co-expressed with ␣7
To detect the presence of surface receptors, transfected cells were labeled by Alexa Fluor 488/␣-BTX, a selective ␣7 antagonist, and examined by confocal microscopy. The confocal images showed high levels of expression of ␣7 but no binding of Alexa Fluor 488/␣-BTX to cells expressing dup␣7 as described before (21) (Fig. 2). Analysis of the membrane fluorescence intensity revealed reduced fluorescence in cells that co-expressed dup␣7 with ␣7 (␣7:dup␣7 1:3 cDNA subunit ratio) with respect to those expressing ␣7 (Fig. 2).

dup␣7 does not mediate functional responses elicited by ␣7 orthosteric or allosteric agonists
To examine function, ␣7 or dup␣7 were expressed in BOSC-23 cells and examined by single-channel recordings. In the presence of 100 -500 M ACh, ␣7 exhibits single brief openings (ϳ0.25 ms) flanked by long closings, or less often, few openings in quick succession, known as bursts (Fig. 3A dup␣7 subunit function (5,12,31 Fig. 3B), in agreement with the lack of macroscopic responses previously reported (21,22). These experiments were performed in parallel with single-channel recordings from cells of the same batch transfected under identical conditions with ␣7 cDNA to confirm successful transfection and receptor expression. As an additional control, we performed recordings from cells transfected only with Ric-3 and NACHO cDNAs and showed no channel activity elicited by ACh and PNU-120596 (n ϭ 8). Because dup␣7 lacks loops A and D of the agonist-binding site we sought to explore activation by an ␣7 allosteric agonist, 4BP-TQS, which binds to the transmembrane region that is conserved between ␣7 and dup␣7 (36). The underlying hypothesis is that if the absence of loops A and D of the ACh-binding site were the cause of the lack of response, this should be over-   Functional ␣7/dup␣7 heteromeric receptors In principle, possible co-assembly of ␣7 and dup␣7 could be determined by co-expressing both subunits. However, singlechannel activity derived from the homomeric ␣7 will probably be the predominant one and functional individual heteromeric arrangements will be indistinguishable. What is needed is a means to unequivocally distinguish functional heteromeric receptors and to directly associate channel openings to heteromeric ␣7/dup␣7 receptors of a given stoichiometry. Thus, to determine whether the co-assembly of dup␣7 and ␣7 subunits leads to functional receptors and to establish the stoichiometry of the functional heteromeric arrangements, we applied the electrical fingerprinting strategy. The strategy is based on the combined expression of ␣7 with an ␣7 subunit mutant that contains three arginine substitutions at the intracellular M3-M4 loop region (␣7LC for ␣7 low conductance, Fig. 4, top). Although ␣7LC receptors are functional as evidenced by macroscopic current recordings, single channels cannot be detected because the amplitude is reduced to undetectable levels (12, 24 -26, 37). Because of the brief duration of ␣7 channels, opening events cannot be fully resolved due to filter bandwidth limitations. Thus, the strategy needs to be performed in the presence of a modulator, here PNU-120596, which by increasing open channel lifetime allows accurate measurements of channel amplitude (35). Recordings of ␣7 in the presence of 100 M ACh and 1 M PNU-120596 showed a homogeneous amplitude population of 9.8 Ϯ 1.7 pA (Ϫ70 mV membrane potential; n ϭ 5 patches from 3 different cell transfections) for both individual opening events and clusters (Fig. 4A, Table 1). Under the same recording conditions, single channel openings from ␣7LC receptors were not detected due to their low amplitude (Fig.  4B). When ␣7LC was expressed together with ␣7 (1:4 for ␣7:␣7LC cDNA subunit ratio), instead of the homogenous amplitude population of clusters detected for ␣7 alone, clusters of different amplitudes were observed (Fig. 4C). Clusters can be grouped into different amplitude classes, which can be welldistinguished from the amplitude histograms (Figs. 4C and 5A). Studies from our and other labs have shown that the different amplitude populations report the number of low conductance subunits in each type of pentameric arrangement (12,26,37). Thus, amplitude classes of clusters of ϳ4-, 6-, 8-, and 10-pA (Ϫ70 mV membrane potential) correspond to receptors containing 3, 2, 1, and 0 ␣7LC subunits, respectively ( Fig. 5A) (12).
In previous studies we also analyzed a 2-pA class that corresponds to receptors containing 4 ␣7LC subunits (12). Here, information from this class was obtained using the reverse combination (see below).
Clusters in the presence of ACh and 1 M PNU-120596 were not detected from cells transfected with ␣7LC cDNA (Fig. 4B) (n ϭ 19 of 4 different transfections) (12) or with dup␣7 cDNA (Fig. 4D, n ϭ 23). In contrast, they were detected from cells expressing both subunits (␣7LC:dup␣7, 1:3, 1:4, and 1:6 cDNA subunit ratios) (Fig. 4E). This result unequivocally indicates that dup␣7 assembles with ␣7LC. It is important to note that the frequency of the active patches was markedly lower than that observed for the ␣7LC/␣7 combination. Under similar transfection conditions, 31% of cell patches showed channel activity in cells co-transfected with ␣7LC and dup␣7 cDNAs (53 of 171 patches), whereas this percentage increased to about 80% in cells transfected with ␣7LC and ␣7 cDNA.
The analysis of the amplitudes of clusters obtained from recordings of cells expressing the ␣7LC/dup␣7 combination showed two predominant classes whose mean amplitude values were 4.2 Ϯ 0.3 and 5.8 Ϯ 0.5 pA (Fig. 5B). In these recordings, we did not analyze the lowest 2-pA amplitude class, which would correspond to receptors containing one dup␣7 subunit because channels of this stoichiometry were better detected using the reverse subunit combination (see below). In 27% of patches, a few clusters of higher amplitude were detected whose origin remains unknown. Because dup␣7 conserves the portal amino acid residues responsible for ␣7 conductance, we can infer that the relationship between the mean amplitude of each class and receptor stoichiometry is for dup␣7/␣7LC the same as for ␣7/␣7LC. Thus, we can ensure that the detected 6-and 4-pA amplitude classes in the ␣7LC/dup␣7 combination correspond to heteromeric receptors containing three and two dup␣7 subunits, respectively (Figs. 4E and 5B).
To determine whether heteromeric receptors containing only one dup␣7 subunit are functional and to further confirm that the three portal amino acid residues in dup␣7 govern chan- Table 1 Kinetic properties of ␣7 and ␣7/dup␣7 receptors Cells expressing the specified subunit combination were used for single-channel recordings. Channels were activated by 100 M ACh in the absence or presence of 1 M PNU-120596. For ␣7, a single ϳ10-pA class is detected. The mean durations of open ( o ) and clusters ( cluster ) were obtained from the corresponding histograms. Channel events from (␣7LC) 2 (dup␣7) 3 and (␣7LC) 2 (␣7) 3 heteromeric receptors correspond to the 6-pA amplitude class recorded from cells transfected with dup␣7 and ␣7LC or ␣7 and ␣7LC, respectively (Fig. 4E). The differences of durations among all receptors were not statistically significant (p ϭ 0.23 for o and 0.62 for cluster , one-way analysis of variance).

Subunit combination
Receptor

dup␣7 subunit function
nel amplitude, we introduced the triple mutation (RRR) in dup␣7 to generate a low conductance dup␣7 subunit (dup␣7LC). We next transfected BOSC-23 cells with the ␣7/dup␣7LC combination (1:8 cDNA subunit ratio) and recorded single-channel currents in the presence of 100 M ACh and 1 M PNU-120596. The amplitude of the majority of clusters in all active patches was ϳ10 pA, which corresponds to that of homomeric ␣7 receptors, indicating the prevalence of this receptor over heteromers. However, clusters of lower amplitude, which were not detected in cells transfected with ␣7 alone, were detected. The analysis showed clusters of 10-, 8-, and 6-pA amplitude classes (Fig. 4F), which correspond to receptors containing zero, one, and two dup␣7 subunits, respectively. In these experiments, we did not analyze amplitude classes lower than 6 pA, because the corresponding populations were well-detected with the reverse combination (␣7LC/dup␣7) (Fig. 4E). Thus, the application of the electrical fingerprinting strategy revealed that dup␣7 can assemble with ␣7 forming functional heteromeric receptors containing one, two, or three dup␣7 subunits.

Arrangement of (␣7) 2 (dup␣7) 3 receptors
The ACh-binding site is located at subunit interfaces (Fig.  6A). A conserved tyrosine (Tyr-93) located in loop A of the principal face of the binding site has been shown to be essential for channel activation (38). Because dup␣7 lacks Tyr-93 (see below), it is likely that it cannot contribute to an activable principal face. We explored if it contributes to the complementary face of the binding site although it lacks loop D. Two possible pentameric arrangements containing three dup␣7 and two ␣7 subunits may be formed, depending on whether the two ␣7  Each plot corresponds to a single recording, and each point, to a single cluster. The number of amplitude classes was determined by the X-means algorithm included in the QuB software.

dup␣7 subunit function
subunits are adjacent or not (Fig. 6B). If the two ␣7 subunits were not consecutive, activation would occur through the ␣7/dup␣7 interface where dup␣7 should provide an activable complementary binding-site face. To test this hypothesis, we co-expressed dup␣7 with an ␣7 subunit carrying a mutation at loop D of the complementary face of the binding site (␣7W55T). We have previously shown that ACh does not elicit either single-channel nor macroscopic currents from cells expressing ␣7W55T receptors (Fig. 6C) (12,24). We did not detect any channel activity elicited by 100 M ACh in the presence of 1 M PNU-120596 from cells transfected with the ␣7W55T/dup␣7 combination (subunit ratios 1:3 and 3:1, n ϭ 10 patches from three different cell transfections) (Fig. 6C). These results indicate that the complementary face of the binding site has to be provided also by the ␣7 subunit to allow activation. Thus, activation of the ␣7/dup␣7 heteromeric receptor occurs through the ␣7/␣7 interfacial-binding site. In consequence, in the pentameric (␣7) 2 (dup␣7) 3 arrangement, the two ␣7 subunits are located consecutively.
Finally, we explored if channel kinetics elicited by 100 M ACh in the presence of 1 M PNU-120596 of (␣7) 2 (dup␣7) 3 receptors are different to those of ␣7. To this end, we analyzed the 6-pA amplitude class of channels recorded from cells expressing ␣7LC and dup␣7, which corresponds to receptors containing three dup␣7 subunits, (␣7LC) 2 (dup␣7) 3 (Figs. 4E and 5B). We found that the mean duration of the slowest open component and the mean cluster duration were not statistically significantly different from those of ␣7 or (␣7LC) 2 (␣7) 3 obtained from the 6-pA population of the ␣7LC/␣7 combination (Table 1). This analysis indicates that the truncated subunit does not alter the channel kinetics of potentiated receptors.

Discussion
The expression and function of human ␣7 can be regulated at different stages and by different mechanisms, such as gene regulation through transcriptional mechanisms (6), co-expression of chaperone proteins (39,40), receptor up-regulation (41), interaction with intracellular proteins (42), and allosteric modulation by endogenous compounds (43). Another mechanism of potential modulation involves the partial duplication of the parent gene, an event that is evolutionary new and human specific (6,44). The mechanism underlying such modulation and the physiological role of the truncated ␣7 subunit remain unknown.
To explore if the truncated subunit resulting from gene duplication, dup␣7, modulates ␣7 function, we generated a dup␣7 cDNA, expressed it on mammalian cells, and deciphered receptor function by single-channel recordings. By using a novel electrophysiological strategy, our results revealed that: (i) dup␣7 alone does not form functional ion channels; (ii) dup␣7 subunit can assemble with ␣7 forming a variety of heteromeric ␣7/dup␣7 receptors; (iii) for functional heteromeric ␣7/dup␣7 receptors, at least two ␣7 subunits are required; (iv) activation of heteromeric receptors requires an ␣7/␣7 interfacial-binding site; and (v) the kinetic signature of potentiated ␣7 receptors is not affected by dup␣7.
Western blotting using an antibody against the ␣7 intracellular loop, which is conserved between ␣7 and dup␣7, showed that dup␣7 cDNA is well-translated in BOSC-23 cells. No ␣-BTX binding was detected in cells transfected with dup␣7 cDNA, in agreement with previous results obtained in oocytes (21). Although no ␣-BTX binding was observed, heterologous expression of dup␣7 homomers in the rat cell line GH4C1 as well as in oocytes was detected using an ␣7 antibody (21). However, it was described that such expression appeared to be at a more inner location than that of ␣7, probably within the endoplasmic reticulum (21). Although the probable absence of a signal peptide of the truncated protein suggests a subcellular localization, whether dup␣7 homomers are present at the surface remains undefined. It is here important to note that a specific dup␣7 antibody, which would facilitate its detection, is still not available.
It has been previously shown that the presence of dup␣7 reduced significantly the number of ␣7 receptors in oocytes (21) but not in neuronal cells (22). However, in the latter system poor dup␣7 translation was reported. In our system, overexpression of dup␣7 (3-fold higher cDNA amount than ␣7 cDNA) reduced ␣-BTX binding at the membrane level. However, we acknowledge that the level of reduction of fluorescence mediated by the presence of dup␣7 may not be accurately determined due to possible bias introduced during the selection of fluorescent cells. In addition, only cells showing membrane fluorescence were analyzed. Assuming that translation and assembly are similar between ␣7 and dup␣7 subunits, the binomial distribution indicates that 23% of the receptors should be dup␣7 homomers and 39.5% should contain only one ␣7 subunit in cells transfected with 1:3 ␣7:dup␣7 cDNA subunit ratio. Under this scenario, an important reduction in ␣-BTX binding should occur because more than 60% of the receptors would not It is important to note that although we used Ric-3 and NACHO as chaperones, their actions on dup␣7 as well as the most appropriate ␣7 subunit:chaperone ratio remain unknown. Also, the expression of dup␣7 appears to depend on the cell type, i.e. in neurons the dup␣7:␣7 protein ratio is opposite to that in immune cells, dup␣7 being the major product in the latter cells (17). Thus, this negative modulation of dup␣7 on ␣7 expression observed in the heterologous system might not be straightforward extrapolated to native systems.
Cells expressing only dup␣7 did not show any detectable single-channel activity elicited by ACh in the presence of the potent PAM PNU-120596. This result supports the consensus that dup␣7 does not form functional receptors in oocytes and mammalian cells (21,22). A hypothesis in support of the absence of response is the lack of an intact ACh-binding site. To overcome the lack of an intact orthosteric agonist-binding site, we used the ␣7 allosteric ligand, 4BP-TQS, which binds to the transmembrane region that is conserved between ␣7 and dup␣7 (36). This ligand mediated strong responses in ␣7 but did not elicit neither macroscopic nor single-channel currents in cells expressing dup␣7. These results confirm the absence of functional dup␣7 receptors, which could arise from either the absence of dup␣7 homomeric receptors in the membrane or from their inability to function as an ion channel.
To gain further insights into why the truncated subunit cannot form functional channels, we modeled dup␣7 using I-TASSER server (45) (Fig. 7). Interestingly, the FAM7A peptide superimposed with loop A in the ␣7 structural model. However, in dup␣7 this region lacks Tyr-93, which is required for ␣7 activation (38). Also, the model shows the lack of loop D at the complementary face, which carries Trp-55 that is important for ␣7 function (25). Moreover, the lack of functional dup␣7 channels is expected because this subunit also lacks the ␤1␤2 loop (Fig. 7), which is located at the coupling region and is required for channel opening (7)(8)(9).
The electrical fingerprinting strategy has been extensively used for determining functional stoichiometry of ␣7-containing receptors (12,26,35,37). This strategy requires the accurate measurement of channel amplitude, which acts as the reporter of the stoichiometry of each receptor that originated a single opening event or cluster. Given the brief open-channel lifetime of ␣7, the strategy needs to be performed in the presence of a PAM that by increasing the open duration allows full channel amplitude resolution (12,35). We chose PNU-120596 because it binds to a site that is conserved between ␣7 and dup␣7 and at the same time it greatly increases opening probability thus facilitating functional detection of low-expressing receptors (46).
The ␣7LC carries a triple mutation in determinants of channel conductance, which are located at the loop between M3 and M4 transmembrane segments at the intracellular end of the ion channel (7,47,48). The triple mutation does not affect singlechannel kinetics and only decreases ␣7 channel amplitude to undetectable levels (12,24). As described in previous studies (12,26,37), when ␣7LC was co-expressed with ␣7, multiple and discrete amplitude classes were detected, each one corresponding to a different population of receptors with a fixed number of low conductance subunits. When instead of ␣7, dup␣7 was expressed with ␣7LC, clusters of different amplitudes activated by ACh in the presence of PNU-120596 were detected. This result unequivocally indicates the presence of surface ␣7LC/ dup␣7 functional receptors because no channel activity was detected with either of the two individual subunits. The application of the electrical fingerprinting strategy revealed that dup␣7 can assemble with ␣7 forming functional heteromeric receptors composed of one, two, or three dup␣7 subunits.
Because loop A with its key tyrosine (Tyr-93) is missing in dup␣7 we infer that in the ␣7/dup␣7 heteromers the ␣7 subunit should provide the principal face of the binding site. The lack  Figure 7. Superimposed molecular models of dup␣7 and ␣7. A, structural alignment of extracellular domains of two adjacent ␣7/␣7 and dup␣7/dup␣7 subunits. Human ␣7 structural model was created by homology modeling based on the structure of the ␣7-AChBP chimera (PDB code 5AFM) and the 3D model of dup␣7 was generated by the I-TASSER server (see "Experimental procedures"). The ␣7 subunits are shown in gray. In dup␣7 subunits the region corresponding to FAM7A is shown in red and that corresponding to ␣7 in blue. Binding and interface loops present in both molecules are indicated with blue letters and those present in ␣7 but absent in dup␣7 with black letters. B, alignment of the ␣7 and dup␣7 sequences (accessions numbers CAD88995 and NP_647536). The ␣7 sequence does not include the signal peptide. The sequences are identical after amino acid residue 95 of ␣7. dup␣7 sequence corresponding to the FAM7A region is in red. Residues for the six binding loops (A-F) and loops at the coupling region are indicated with black and gray lines, respectively. Aromatic residues reported as essential for ␣7 agonist response are in gray boxes.

DLQMQEADISGYIPNGEWDLVGIPGKRSERFYECCKEPYPDVTFTVTMRRR DLQMQEADISGYIPNGEWDLVGIPGKRSERFYECCKEPYPDVTFTVTMRRR
dup␣7 subunit function of functional responses from cells expressing dup␣7 and ␣7W55T, which does not contain a functional complementary face of the binding site, indicates that this face must also be provided by ␣7. Thus, in (␣7) 2 (dup␣7) 3 receptors, the two ␣7 subunits are located consecutively and activation takes place through agonist binding at their interface. The fact that (␣7) 2 (dup␣7) 3 can be activated despite carrying only one intact agonist-binding site is in close agreement with our previous results showing that only one functional ACh-binding site is sufficient for ␣7 activation and that the four additional sites increase ACh sensitivity (12). It also agrees with reports of ␣7␤2 receptors showing that activation of this heteromeric receptor takes place through the ␣7/␣7 interface (49,50). Thus, it appears that in ␣7-heteromeric receptors at least one ␣7/␣7 interfacial-binding site is required for function. We also showed that the mean durations of channel openings and clusters of (␣7) 2 (dup␣7) 3 elicited by ACh in the presence of PNU-120596 are identical to those of ␣7, indicating that, at least in PNU-120596-potentiated receptors, the kinetics are not affected by dup␣7. Unfortunately, this strategy cannot be performed in the absence of a potentiator due to the lack of full amplitude resolution and the low frequency of opening events (35).
Overall, our electrophysiological results predict that dup␣7 will functionally operate as a negative modulator of ␣7 activity. Heteromers containing four dup␣7 subunits are nonfunctional and those with three or less dup␣7 subunits, despite being functional, have reduced ACh sensitivity due to the reduced number of active ACh-binding sites (12,25). An additional negative modulatory action of dup␣7 might be associated with the decrease of the number of surface ␣7 receptors (21). However, this may differ between native and heterologous systems due to differences in gene expression, cell-surface translocation, channel assembly, or chaperones. Thus, our findings encourage to explore the assembly of heteromers in different human tissues.
Our study has been focused on ionotropic responses. However, ␣7 has been shown to act as a dual ionotropic/metabotropic receptor (5,42). Considering that the ␣7 channel-independent signal transduction is important in anti-inflammatory responses and that immune cells show high expression of dup␣7 (17), it will be interesting to determine whether the metabotropic activity differs between homomeric and ␣7/dup␣7 heteromeric receptors. From a molecular point of view, our findings provide novel information regarding ␣7 unique activation, and from a physiological point of view, they help to reveal the still unknown impact of the human-specific truncated subunit on ␣7 function.

Cloning of dup␣7 cDNA
The full-length CHRFAM7A (variant 1: NM_139320.1) cDNA that encodes a 27-amino acid terminus corresponding to the FAM7A sequence (NH 2 -MQKYCIYQHFQFQLLIQHL-WIAANCDI) and thereafter ␣7 sequence starting at ADERFDA, which corresponds to the end of loop A of the binding site, was synthesized de novo (Biomatik, USA) with appropriate flanking restriction sites, XbaI and HindIII, for subclon-ing into the pUC19 plasmid to generate pUC19-dup␣7. The dup␣7 cDNA was excised from pUC19 -dup␣7 and subcloned into the cytomegalovirus-based expression vector pRBG4 (51). After cloning, the sequence of dup␣7 cloned in pRBG4 was confirmed by DNA sequencing using capillary electrophoresis (Instituto de Biotecnología CICCVyA, INTA, Argentina).

Cell expression
BOSC-23 cells, derived from HEK-293 cells (51), were transfected by the calcium phosphate procedure with dup␣7 and/or human ␣7 cDNA (also subcloned in pRBG4 vector), essentially as described previously (52,53). For the ␣7LC/dup␣7 and ␣7/dup␣7LC combinations the cDNA subunit molar ratios ranged between 1:3 and 1:10 to ensure the incorporation of dup␣7 into the heteromeric receptor. Although the cDNA subunit ratio is not directly proportional to the subunit stoichiometry of the final receptor, an excess of one subunit over the other is expected to compensate for less efficient translation or assembly and to bias the receptor population toward pentamers with an excess of the surplus subunit (24,54). Plasmids harboring cDNAs of the ␣7 chaperone proteins Ric-3 and NACHO were incorporated in all transfections (39,55).

Confocal fluorescence microscopy
Cells were plated on 12-mm glass coverslips in 35-mm dishes. They were transfected with ␣7 cDNA (0.3 g), dup␣7 cDNA (0.9 g), or with the combination (1:3) of ␣7 (0.3 g) and dup␣7 (0.9 g) cDNAs. Ric-3 (1 g) and NACHO (1 g) cDNAs were included during the transfection in all conditions. Mock transfected cells correspond to cells transfected with irrelevant plasmid DNA (1.2 g). The total amount of DNA per transfection was normalized with the addition of irrelevant plasmid DNA. After 72 h, cell-surface receptor labeling was carried out by incubating with ␣-BTX Alexa Fluor 488 conjugate (Molecular Probes) at a final concentration of 1 g/ml for 1 h in chilled Dulbecco's modified Eagle's medium. Cells were then fixed with 4% paraformaldehyde. Cultures were then analyzed by phase and fluorescence microscopy, using a Nikon Eclipse E600 microscope and by laser scanning confocal microscopy (LSCM; Leica DMIRE2) with a ϫ20 objective. Fluorescence intensity, expressed as arbitrary units, was measured after manually outlining regions of interest with the software ImageJ (National Institutes of Health, Bethesda, MD). The maximum fluorescence intensity of a given region of interest was measured within the ␣-BTX-positive region of the cell surface, and the maximum fluorescence intensity of an area of the same size positioned over a ␣-BTX-negative region outside the cell was subtracted. The average fluorescence intensity values were calculated for randomly chosen cells for each experimental condition.

Single-channel recordings
Single-channel recordings were obtained in the cell-attached patch configuration (31). The bath and pipette solutions contained 142 mM KCl, 5.4 mM NaCl, 1.8 mM CaCl 2 , 1.7 mM MgCl 2 , and 10 mM HEPES, pH 7.4. For potentiation, 1 M PNU-120596 was added to the pipette solution with ACh (12).
Single-channel currents were digitized at 5-10-s intervals, low-pass filtered at a cut-off frequency of 10 kHz using an Axopatch 200B patch clamp amplifier (Molecular Devices Corp.). Single-channel events were idealized by the half-amplitude threshold criterion using the program QuB 2.0.0.28 with a digital low-pass filter at 9 kHz. A filter of 3 kHz was used in recordings with PNU-120596 to facilitate the analysis. The open and closed time histograms obtained from idealization were fitted by the maximum interval likelihood function in QuB (56,57), with a dead time of 0.03-0.1 ms. This analysis was performed on the basis of a kinetic model whose resulting probability density function curves properly fit the histograms following the maximum likelihood criteria. For ␣7, this analysis was done by sequentially adding an open and/or closed state to a starting C 7 O model to properly fit the corresponding histograms. Final models contained 5-6 closed states and 3-4 open states for ␣7 in the presence of ACh plus PNU-120596, or three closed states and 1-2 open states for ␣7 in the presence of ACh alone (31,32,34).
Clusters were identified as a series of closely separated openings preceded and followed by closings longer than a critical duration. Different critical closed times were calculated by maximum interval likelihood between each closed component. Critical times between the third and fourth closed components for ␣7 in the presence of PNU-120596 (ϳ30 to 60 ms) were selected in QuB to chop the idealized data and create a subdata set that only contained clusters to define mean cluster duration.

Electrical fingerprinting strategy
To define amplitude classes from receptors generated by coexpression of high and low conductance subunits, all clusters were selected regardless of their current amplitudes. Amplitude histograms were then constructed, and the different amplitude classes were distinguished. The number of amplitude classes and the mean cluster duration for each class were determined by the X-means algorithm included in the QuB software. Although up to 4 different amplitude classes were detected using the X-means algorithm, not all recordings contain events of all classes. Open time histograms were determined as described above for a selected amplitude class.
Results with QuB analysis were similar to those obtained with TAC and TAC fit programs (Bruxton Corp., Seattle, WA) as described before (12,24,25). Briefly, events were detected by the half-amplitude threshold criterion using the program TAC. To define amplitude classes, analysis was performed by tracking events regardless of current amplitude. Amplitude histograms were then constructed and fitted by TAC fit. The different amplitude classes were distinguished by this way in experiments shown in Fig. 4.

Statistical analysis
Unless otherwise noted, data were presented as mean Ϯ S.D. Statistical comparisons were done using pairwise t test or oneway analysis of variance with GraphPad Prism (GraphPad Software Inc.). Statistically significant differences were established at p values Ͻ 0.05.

Molecular modeling
A homology model for the extracellular region of human ␣7 receptor was created based on the structure of the ␣7-AChBP chimera (PDB code 5AFM). The amino acid sequence for the human ␣7 subunit (accession number: CAD88995.1) was aligned and modeled using MODELLER 9v8 (58). Ten models were generated; of these, the one with the lowest energy and the smallest percentage of amino acid residues in the disallowed region of the Ramachandran plot was selected. To obtain a 3D model of dup␣7 we used the I-TASSER server (45). This server performs structure and function prediction for a query amino acid sequence by a combination of homology modeling, threading, and ab initio modeling. Five models were generated, and the best model was selected on the base of the C-score. Structure analysis and figures were generated using Discovery Studio Visualizer v4.5 suite (Dassault Systemes BIOVIA, San Diego).