Function of Partially Duplicated Human α7 Nicotinic Receptor Subunit CHRFAM7A Gene

The neuronal α7 nicotinic receptor subunit gene (CHRNA7) is partially duplicated in the human genome forming a hybrid gene (CHRFAM7A) with the novel FAM7A gene. The hybrid gene transcript, dupα7, has been identified in brain, immune cells, and the HL-60 cell line, although its translation and function are still unknown. In this study, dupα7 cDNA has been cloned and expressed in GH4C1 cells and Xenopus oocytes to study the pattern and functional role of the expressed protein. Our results reveal that dupα7 transcript was natively translated in HL-60 cells and heterologously expressed in GH4C1 cells and oocytes. Injection of dupα7 mRNA into oocytes failed to generate functional receptors, but when co-injected with α7 mRNA at α7/dupα7 ratios of 5:1, 2:1, 1:1, 1:5, and 1:10, it reduced the nicotine-elicited α7 current generated in control oocytes (α7 alone) by 26, 53, 75, 93, and 94%, respectively. This effect is mainly due to a reduction in the number of functional α7 receptors reaching the oocyte membrane, as deduced from α-bungarotoxin binding and fluorescent confocal assays. Two additional findings open the possibility that the dominant negative effect of dupα7 on α7 receptor activity observed in vitro could be extrapolated to in vivo situations. (i) Compared with α7 mRNA, basal dupα7 mRNA levels are substantial in human cerebral cortex and higher in macrophages. (ii) dupα7 mRNA levels in macrophages are down-regulated by IL-1β, LPS, and nicotine. Thus, dupα7 could modulate α7 receptor-mediated synaptic transmission and cholinergic anti-inflammatory response.

Neuronal ␣7 nicotinic acetylcholine receptors (␣7 nAChRs) 4 are widely expressed in the central and peripheral nervous systems. In neurons, homomeric ␣7 nAChRs, composed of five ␣7 subunits, modulate neurotransmitter release in presynaptic nerve terminals and induce excitatory impulses in postsynaptic neurons (1)(2)(3)(4). Signaling through ␣7 nAChRs in the central nervous system has been associated with neuronal plasticity and cell survival (5)(6)(7), although impaired activity of this receptor has been implicated in the pathogenesis of schizophrenia, Alzheimer disease, and depression (8 -12). The presence of ␣7 nAChRs has also been reported in non-neuronal cells such as vascular and brain endothelial cells, bronchial epithelial cells, keratinocytes, astrocytes, synoviocytes, thymocytes, lymphocytes, bone marrow cells, monocytes, macrophages, microglia, and astrocytes (see Ref. 13 and the references therein). Interestingly, the ␣7 nAChR expressed in macrophages (and probably in other immune cells) is essential for vagus nerve regulation of acute proinflammatory cytokine release during systemic inflammatory response (see Ref. 14 and references therein). The ␣7 nAChR is a target for natural and synthetic ligands; however, little is known about endogenous receptor-regulating molecules.
The ␣7 nicotinic subunit encoded by gene CHRNA7 is located on the long arm of chromosome 15 (15q13-q14). A hybrid gene (CHRFAM7A) resulting from a fusion of a partial duplication of CHRNA7 with FAM7A gene was identified at 1.6 Mb from CHRNA7 toward the centromeric region (15,16). CHRNA7 and CHRFAM7A are highly homologous (Ͼ99%) from exon 5 to the 3Ј-UTR region. In contrast, the hybrid has replaced exons 1-4 of the original CHRNA7 with the exons D, C, B, and A of FAM7A, inserted in an Alu sequence of intron 4 of CHRNA7, 700 bp upstream from exon 5. Moreover, CHRFAM7A is polymorphic with a few rare individuals who completely lack a copy of the gene. Most people (Ͼ95%) have one or two CHRFAM7A copies, and in some cases, there is a 2-bp deletion in exon 6 (16,17). The acquisition of this duplication seems to be a recent evolutionary event because CHRFAM7A only appears in humans and not in other higher primates (18). The above findings would indicate that the hybrid gene could confer an evolutionary advantage on the genotype with the duplication.
To date, the possible functional significance of this CHRNA7 duplication is unknown, although there are conflicting reports in the literature suggesting an association between CHRFAM7A polymorphisms and certain psychiatric and neurological disorders, such as schizophrenia, bipolar depression, Alzheimer disease, dementia with Lewy bodies, or Pick disease (19 -24). The CHRFAM7A transcript, dup␣7, has been identified in hippocampus, cortex, corpus callosum, thalamus, putamen, caudate nucleus, and cerebellum (15,(25)(26) and in peripheral blood mononuclear cells (PBMC), lymphocytes, synoviocytes, as well as HL-60 cells (25)(26)(27)(28). Despite its wide distribution, to date there is no experimental evidence demonstrating that this transcript is translated and what the possible functional role of the resulting protein might be.
This study seeks to shed some light on the last two questions. Using GH4C1 cells and oocytes, we have studied the function of heterologously expressed dup␣7 protein, and, to study the native transcript, we use HL-60 cells, human CTX, and human macrophages. A set of experimental approaches, including molecular biology and confocal images of labeled receptors, combined with pharmacological and electrophysiological techniques, has been used throughout this study.

EXPERIMENTAL PROCEDURES
Cell Lines and Oocytes-The rat pituitary-derived GH4C1 cells were grown in DMEM containing 10% fetal calf serum, and the human acute promyeloid leukemic HL-60 cell line in RPMI 1640 medium supplemented with 10% fetal bovine serum. In both cases, 100 units⅐ml Ϫ1 of penicillin G sodium and 100 g⅐ml Ϫ1 of streptomycin sulfate were added. Both cell types were maintained at a density of 1 ϫ 10 6 to 2 ϫ 10 6 cells/ml at 37°C in a humidified 5% CO 2 atmosphere. Mature female Xenopus laevis frogs were obtained from a commercial supplier (Xenopus Express, Haute-Loire, France). Techniques for oocyte isolation have been described by our group previously (29 -32).
Isolation of Human Monocytes and Differentiation to Macrophages-The PBMCs were isolated from blood buffy coats from healthy individual donors, and the monocyte-enriched cell population was then purified from PBMC by a double density gradient protocol as described previously (33). Monocytes were collected, washed, resuspended in complete RPMI 1640 medium, and seeded either onto coverslips in 24well tissue culture plates (5 ϫ 10 5 cells per well) or onto 60-mm Petri dishes (10 6 cells⅐ml Ϫ1 ). Cells were allowed to differentiate to macrophages (MØ) for 7-9 days in the presence of macrophage colony-stimulating factor (2 ng⅐ml Ϫ1 ) in complete culture medium.
Immunofluorescent Staining and Confocal Microscopy-GH4C1 cells were electroporated with plasmids dup␣7⅐ pcDNA3, ␣7⅐pcDNA3, or pcDNA3. They were then plated on 12-mm glass coverslips for 48 h before processing. Cells were fixed with 4% paraformaldehyde, permeabilized with 0.25% Triton X-100, and stained overnight with the monoclonal anti-␣7 subunit antibody (Mab306, Sigma, 1:3000) in PBS with 5% normal goat serum at 4°C, followed by incubation with AlexaFluor 488 goat anti-mouse IgG (1:400) for 2 h at room temperature. All images were captured with a Leica TCS SP2 spectral confocal laser scanning microscope as reported previously (31,32). Fluorescence intensity, expressed as arbitrary units, was measured in a narrow region of interest (ROI) located between the plasma membrane and the cell nucleus. In other cases, fluorescence intensity was measured along the x axis of the entire cell body.
Techniques for mRNA injection and immunodetection of foreign proteins expressed in intact injected oocytes have been described elsewhere (31,32). Five days after injection, protein expression was analyzed using the same antibodies and dilutions as for GH4C1 cells. Oocytes were mounted in glycerol/buffer (70:30) under a glass coverslip on a glass dualwell slide. The slide was fastened upside down on the stage of the Leica confocal microscope and visualized with a ϫ20 lens. The fluorescent signal was determined as arbitrary units in an ROI located on the surface of the animal pole of the oocyte.
Electrophysiological Recordings and Surface Receptor Binding Assay-Electrophysiological recordings of foreign proteins expressed in oocytes have been described in detail by our group elsewhere (29,32,35). The agonist concentration eliciting half-maximal current (EC 50 ) and the Hill coefficient values were estimated through nonlinear regression analysis using the four-parameter logistic equation of the GraphPad Prism software. Total surface expression of ␣Bgtx-binding sites on the surface of mRNA-injected oocytes was assayed with 125 I-␣ Bgtx (Amersham Biosciences), as described elsewhere (36). Briefly, 5 days after injection, oocytes placed on 24-well plates were preincubated for 15 min with Ringer's solution containing fetal calf serum (6%). Then they were incubated for 2 h at 18°C in a final volume of 300 l of the above solution containing increasing concentrations (1-10 nM) of 125 I-␣ Bgtx. After incubation, labeled medium was removed and oocytes washed several times with Ringer's solution before counting total bound radioactivity in a Beckman Coulter ␥ counter. Nonspecific binding was determined in noninjected oocytes incubated under the same conditions; this value was subtracted from total binding to yield specific binding.
Northern Blot Analysis-Poly(A) ϩ RNAs (4 g/lane) isolated from human CTX and from HL-60 cells were separated by 1% formaldehyde/agarose gel electrophoresis, transferred onto a nylon blotting membrane (Schleicher & Schuell), and hybridized at 68°C overnight with the 32 P-labeled probe (781 bp) obtained by digestion of dup␣7⅐pMOS-Blue with SpeI and SmaI. The resulting fragment starts in the ATG of exon B and extends to the 51st nucleotide of exon 10; thus, it recognizes the dup␣7 and ␣7 transcripts.
Isolation of Polysome-bound mRNA and Real Time Quantitative PCR (Q-PCR)-Polysome-bound mRNA from HL-60 cells was obtained as described by del Prete et al. (37), except that cells were pretreated with 0.1 mg⅐ml Ϫ1 cycloheximide for 10 min at 37°C. In the conditions used, ribonucleoprotein particles, isolated ribosomal subunits, and monosomes remain in fractions 1-10 (from the top) whereas polysomes are recovered in fractions 11-20. After this, monosomic and polysome fractions were regrouped into two pools, and 1 aliquot (10 l, diluted to 100 ng⅐l Ϫ1 ) from each pool was used for Q-PCR mRNA quantification. We added 10 ng of in vitro transcribed luciferase mRNA (from pGL3-control plasmid) to each aliquot as an endogenous marker for quantification. The level of gene expression in human CTX and human MØ was also assessed by Q-PCR for reverse-transcribed mRNA or total RNA, respectively. A combination of TaqMan and SYBR Green-based assays for amplicon detection on the ABI Prism 7900 device was used. The following primers were used: for dup␣7, forward 5Ј-CAATTGCTAATCCAGCATTTGT-3Ј (position 603-624 of sequence in NM_139320) and reverse 5Ј-CCCAGAAGAATTCACCAACACG-3Ј (position 704 -683 in NM_139320); for ␣7, Hs_CHRNA7_2_SG, QuantiTech Primer Assay, QT01681211 (Qiagen); for TBP (encoding for the TATA-binding protein), forward 5Ј-CGGTTTGCTGCG-GTAATCA-3Ј and reverse 5Ј-TGTTGGTGGGTGAGCAC-AAG-3Ј; for luciferase, forward 5Ј-TGGAAGACGCCAAAA-ACATAAAG-3Ј and reverse 5Ј-AGCAATTGTTCCAGGAA-CCAGGGC-3Ј. For dup␣7 and ␣7 mRNA quantification in MØ, TaqMan Hs00415199_m1 and Hs01063372_m1 (Applied Biosystems) were used, respectively. Analysis of the melting curves demonstrated that each pair of primers amplified a single product. Q-PCR for each mRNA was performed in triplicate, and all results were normalized to the expression of 18 S rRNA (Ambion primers). Relative gene expression values were calculated by the comparative ⌬Ct method using the Sequence Detection System 1.2 software (Applied Biosystems).
Chemicals-Unless otherwise indicated, all products were purchased from Sigma. AlexaFluor 488 goat anti-mouse IgG was from Molecular Probes (Eugene, OR), and DMEM and RPMI media were from Invitrogen. The human recombinant IL-1␤ was purchased from PeproTech (London, UK), and PNU120596 was from Tocris Bioscience (Bristol, UK).

RESULTS
Analysis of the dup␣7 Coding Sequence-The cloned fulllength dup␣7 cDNA codes for a 1236-nucleotide sequence similar to the one previously deposited in NCBI and that corresponds to the dup␣7 isoform 1 (accession number NM_139320) with minimal variations. Our sequence contains two neutral transitions, C/T and A/G at positions 654 and 986, respectively, but it retains the two TG nucleotides at positions 497-498 on exon 6. Using the ConPred II software for transmembrane topology prediction, one can observe that the resulting 412-amino acid polypeptide preserves many of the distinctive features of the nAChR subunits, with the N and C termini outside the membrane and the four transmembrane domains (M1-M4) as well as the long intracellular loop between M3 and M4 ( Fig. 2A). However, as has been reported previously (15), dup␣7 protein does lose a substantial part of the long N-terminal domain of ␣7 and thus lacks the recognition sites for ACh and ␣ Bgtx.
Heterologous Expression of dup␣7 Versus ␣7-Expression of foreign proteins was analyzed in GH4C1 cells transfected with plasmids dup␣7⅐pcDNA3 or ␣7⅐pcDNA3 and in oocytes injected with dup␣7 and ␣7 mRNAs. The GH4C1 cell line was selected for two reason. 1) It does not express dup␣7 or ␣7, making visualization of foreign protein expression feasible. 2) It expresses the RIC-3 chaperone (38) making it an excellent model to express functional ␣7 nAChRs. The confocal images obtained in transfected GH4C1 cells stained with Mab306 show a high expression level of both proteins, in contrast to low fluorescence detected in the negative control cells transfected with pcDNA3 ( Fig. 1, A and C). The analysis of the protein expression along the x axis of the cell reveals two expression peaks located in the cell periphery; the ␣7 peak seems to be located more externally than the dup␣7 peak (Fig. 1B).
Noninjected oocytes (control) have low basal fluorescence, which increases markedly at or near the cell surface of the Dominant Negative Effect of dup␣7 on ␣7 Receptor Activity animal hemisphere in dup␣7 or ␣7 mRNA-injected oocytes incubated with the above antibody (Fig. 2B). Nevertheless, the fluorescence signal corresponding to dup␣7 is significantly lower (33 Ϯ 7%, n ϭ 7; p Յ 0.05) than that found in ␣7 mRNA-injected oocytes. As has been reported previously, no intracellular fluorescence was observed in mRNA-injected oocytes under any experimental condition (39). This may be due to the following: 1) the inability of the laser to penetrate far into a large 1-mm cell; 2) the quenching of the fluorescent signal by the oocyte yolk/pigment; or 3) the possibility that intracellular antibody distribution may be too diffuse to be detected.
Dominant Negative Effect of dup␣7 on ␣7 Currents Expressed in Oocytes-Once the heterologous expression of the dup␣7 protein in GH4C1 cells and oocytes was confirmed, we proceeded to study its possible functional role in the latter cell type. The study began with the electrophysiological recording of ACh-elicited currents (I ACh ) in oocytes injected with ␣7 mRNA (2.5 ng/oocyte). In all experiments RIC3 mRNA (2.5 ng) was co-injected with ␣7 mRNA because this treatment almost doubles the amplitude and reproducibility of the re-corded currents. Repeated pulses of ACh (100 M, 1 s) or nicotine (30 M, 1 s) at 1-min intervals, applied to oocytes injected with increasing concentrations of dup␣7 mRNA (from 0.5 to 25 ng/oocyte), failed to produce any response (not shown). In contrast, co-injection of dup␣7 with ␣7 mRNA (2.5 ng) resulted in a concentration-dependent reduction of ␣7 current induced by nicotine (I Nic ) (Fig. 3, A and B). The co-injection of 0.5 ng/oocyte of dup␣7 mRNA (␣7/dup␣7 ratio of 5:1) was sufficient to reduce significantly (26 Ϯ 8%, n ϭ 6; p Յ 0.01) the amplitude of I Nic , and this blockade exceeded 90% when the ␣7/dup␣7 ratio was 1:5 or 1:10.
The next experiments were designed to analyze the nature of the dominant negative effect of dup␣7 on the ␣7 current. One possibility is that dup␣7 is integrated into the pentameric structure of the expressed receptor, changing its pharmacological properties. Another possibility is that dup␣7 interferes with the proper assembly of ␣7 subunits that form mature homopentameric ␣7 nAChR, which is essential for receptor trafficking from endoplasmic reticulum (ER) to the cell membrane. The following experiments sought to explore both possibilities.
dup␣7 Reduces ␣7 Current Amplitude but Not Receptor Response to ACh- Fig. 3 (C and D) shows the concentrationresponse curves to ACh in oocytes injected with 2.5 ng of ␣7 mRNA and the corresponding amount of dup␣7 mRNA according to the indicated ␣7/dup␣7 ratio. Each oocyte was stimulated with successive pulses of increasing concentrations of agonist and, at the end of experiment, with a control pulse of 1 mM ACh that induced the maximum current (I max ). Fig.  3C shows the original traces of I ACh obtained in one ␣7 mRNA-injected oocyte stimulated as described above. As expected, for a given ACh concentration, the higher the proportion of dup␣7 in the mixture of injected mRNA, the lower the I ACh . Thus, the I max for the combinations 1:0, 2:1, 1:1, and 1:5 of ␣7/dup␣7 was 8.4 Ϯ 1.5, 3.9 Ϯ 0.7, 2.1 Ϯ 0.5, and 0.6 Ϯ 0.2 A, respectively. The currents induced by each of the ACh concentrations tested in the oocyte were normalized to the I max . Fig. 3D shows pooled results of normalized currents obtained in several oocytes (n ϭ 5-6) assayed for each condition; each value represents mean Ϯ S.E. The EC 50 and Hill coefficient values are shown in the table below; the analysis of variance (ANOVA) applied to both parameters showed no significant differences between the four groups of oocytes.
dup␣7 Reduces the Number of Functional ␣7 nAChRs in the Oocyte Membrane-To examine whether the dup␣7 effect was due to a reduction in the number of functional ␣7 nAChRs incorporated into the surface of the oocyte, we incubated the oocytes with either the Mab306 antibody, which detects both ␣7 and dup␣7, or with FITC-␣Bgtx, which binds to the extracellular N-terminal region present in the ␣7 subunit alone. Three groups of mRNA-injected oocytes (␣7, dup␣7, or 1:1 ␣7/dup␣7), along with a fourth group without an injection (control) were analyzed. The confocal images in Fig. 4A show the fluorescent signal detected in eight typical intact oocytes from the same donor subjected to one or the other staining process. Fig. 4B shows the average fluorescence values, expressed as a percentage of the control, determined in several oocytes (n ϭ 5-8) used for testing each experimen- tal condition. Although the antibody detected a robust expression of ␣7 nAChRs and a small fluorescent signal generated by dup␣7 in oocytes injected with one or the other mRNA, FITC-␣Bgtx only detected the expression of the first protein. However, regardless of the staining technique used, the co-injection of dup␣7 with ␣7 mRNA significantly reduced the expression of ␣7 nAChRs on the surface of the oocyte. In fact, in oocytes labeled with FITC-␣Bgtx, the ␣7 signal was reduced (Ϸ70%) to levels that were indistinguishable from those in the control. It should be noted that, regardless of the staining process used, confocal images show a polarized distribution of nicotinic subunits on the animal membrane of the oocyte.
Co-expression of dup␣7 with ␣7 Reduces the Number of 125 I-␣ Bgtx-binding Sites on the Oocyte Surface-Two groups of oocytes injected with ␣7 mRNA (control) or with the com-bination 1:4 of ␣7/dup␣7 mRNA were incubated with increasing concentrations of 125 I-␣ Bgtx as described under "Experimental Procedures." Compared with controls, the injection of dup␣7 mRNA reduced markedly and significantly the maximum specific binding of radioligand (B max ), from 8,277 Ϯ 331 to 4,552 Ϯ 288 cpm; p Յ 0.01 (Fig. 4C, bottom  panel). Additionally, the equilibrium dissociation constant (K d ) suffered a slight but significant increase in the presence of dup␣7 (from 1.37 Ϯ 0.17 to 4.28 Ϯ 0.56 nM; p Յ 0.05).
Effect of Positive Allosteric Modulators on I ACh Expressed in Oocytes Injected with Several Combinations of ␣7/dup␣7 mRNA-The following experiments were designed to look for possible pharmacological differences among oocytes injected with different ␣7/dup␣7 mRNA ratios. We used two positive allosteric modulators of ␣7 nAChRs, each one representative of one of the two available allosteric groups. Group I includes agents like 5-hydroxyindole (5HI), which potentiates the peak agonist-evoked response mediated by ␣7 nAChR, but it does not modify the duration of the response (40). Meanwhile, PNU120596, representative of group II, increases the peak agonist-evoked response and markedly prolongs the macro-scopic currents in the presence of agonist, possibly by interfering with the desensitization process (41). The different effects of these two groups of positive allosteric modulators indicate the existence of distinct allosteric binding sites with specificity for either 5HI or PNU120596. The effects of 5HI on I ACh were analyzed in several groups of oocytes injected with different combinations of ␣7/dup␣7 mRNA, using the group of ␣7 mRNA-injected oocytes as reference (Fig. 5, upper left panel). The oocytes were stimulated with successive pulses of ACh (30 M, 1 s), applied at intervals of 1 min. Once the I ACh was stable (control), we incubated the oocytes with 3 mM 5HI (30 s before and during the ACh pulse), evaluating its enhancing effect on control I ACh . Fig. 5 (A and B) shows that 5HI increased the amplitude of control I ACh about 5-fold and that this effect was not significantly different in any oocyte group. Additionally, 5HI did not alter the kinetics of I ACh in any of the oocyte groups as can be deduced after normalization of the original traces obtained in two oocytes injected with either ␣7 or ␣7/dup␣7 (1:1) mRNAs (Fig. 5A,  inset). Oocytes, voltage-clamped at Ϫ70 mV, were stimulated with regular pulses of ACh in the absence or presence of the modulator. The effect of 5HI was evaluated on a single ACh pulse, although in the case of PNU120596 (PNU), the oocyte was stimulated with an initial pulse of ACh (0 ϭ control) followed by six successive ACh pulses in the presence of the modulator. A, original traces of ACh-elicited currents (I ACh ), in the absence or presence of 5HI, obtained in two mRNA-injected oocytes. The left and right records correspond to one ␣7-mRNA and one ␣7/dup␣7-mRNA (1:1)-injected oocyte, respectively. Inset, normalized currents of both oocytes. B, histogram shows the number of times that 5HI increases the control peak I ACh in each group of oocytes. Each bar shows the mean Ϯ S.E. of the number of oocytes in parentheses. Analysis of variance and Dunnett post hoc tests showed no differences between groups of oocytes. C, original traces showing the PNU120596 effect on I ACh over the 1st to 6th pulses of ACh in two mRNA-injected oocytes. The left and right records correspond to one ␣7-mRNA and one ␣7/dup␣7-mRNA (1:1) injected oocyte, respectively. Inset, normalized currents obtained in both oocytes corresponding to the initial (control) and the 3rd pulse of ACh in the presence of PNU. D, graph representing the number of times that P PNU120596 increased the control peak I ACh during the six pulses in each group of oocytes. Each point represents the mean Ϯ S.E. of seven oocytes per group. Data were analyzed using ANOVA and Bonferroni post hoc tests. **, p Յ 0.01; ***, p Յ 0.001; ##, p Յ 0.01 after the comparison of the different oocyte groups as indicated.
A different design was used to evaluate the effect of PNU120596 in oocytes injected with several combinations of ␣7/dup␣7 mRNA (Fig. 5, top right panel). The ␣7 mRNAinjected group was used as reference. The oocytes were stimulated with successive pulses of ACh (30 M, 1 s, at 1-min intervals). Once I ACh was stabilized (control), we perfused 300 nM PNU120596 before and during the ACh pulse, evaluating its enhancing effect on control I ACh . Because there was a progressive positive effect of the modulator on I ACh over successive ACh pulses, we evaluated this effect over six consecutive agonist pulses. Fig. 5 (C and D) shows that the higher the proportion of dup␣7 in the mixture of injected mRNA, the lower the enhancing effect of PNU120596 on peak I ACh . Normalization of the currents (Fig. 5C, inset) shows that the allosteric modulator delays desensitization of I ACh in all the oocyte groups, with no significant differences, probably because of the shortness of the ACh pulse.
Identification of the dup␣7 Transcript in Cerebral Cortex and HL-60 Cells-Although the presence of the dup␣7 transcript has been detected in different brain areas, immune cell types, and the HL-60 cell line by PCR, the size of this mRNA is unknown. Northern blot analysis was performed with mRNAs extracted from human CTX and HL-60 cells because the latter cell type mainly expresses dup␣7 and not ␣7 (25). Fig. 6A shows a well defined band of Ϸ3.9 kb in HL-60 cells, also present in CTX, the latter also containing bands of 5.8, 3.3, 2.1, and 1.4 kb.
dup␣7 mRNA Is Translated in HL-60 Cells-Antibodies that specifically differentiate ␣7 and dup␣7 are not available, so it is still impossible to demonstrate the existence of dup␣7 protein in vitro or in vivo. To solve this difficulty, we have gone a step backwards in the translation process and used an indirect approach, one that looks for dup␣7 mRNA in the polysomal fraction because we know that polysome-bound mRNAs are being translated (42,43). After sucrose gradient centrifugation of an HL-60 cell extract as described under "Experimental Procedures," a bioanalysis assay for each fraction showed that tRNA appears in fractions 1-4, the migration peak for the small 40 S ribosomal subunit (18 S peak) is in fractions 5-6, followed by the migration peak for the large 60 S ribosomal subunit (28 S peak) in fractions 7-10, whereas fractions 11-20 contain the polysomal fraction (28 S and 18 S). All fractions were regrouped into two pools (monosomal and polysomal) as described under "Experimental Procedures." Relative expression of dup␣7, TBP, and luciferase mRNAs was quantified in the two pools. TBP mRNA was used as the positive control because it is constitutively translated in all cell types and therefore must be present in the "polysomal fraction" of HL-60 cells. Fig. 6B shows the relative levels of TBP mRNA and dup␣7 mRNA in monosomal and polysomal pools, expressed as a percentage of the total mRNA in both fractions. The results reveal that, as expected, the TBP transcript was in the heavy sucrose fractions (polysomes), as was the dup␣7 transcript.

Levels of the dup␣7 Transcript in Human Cerebral Cortex and Macrophages and Its Modulation by Different
Treatments-We performed Q-PCR experiments to determine levels of dup␣7 and ␣7 mRNAs in CTX and MØ. The PCR was performed with human CTX mRNA or with total RNA from cultured human MØ using the primers listed under "Experimental Procedures." The results in Fig. 7A show that dup␣7 expression is around 22% of the ␣7 expression in CTX. The opposite happens in MØ, where the ␣7 mRNA levels are ϳ7% of those of dup␣7. Despite this low level of ␣7 mRNA in relation to dup␣7 in MØ, these cells clearly show the expression of functional ␣7 nAChRs, as reflected by confocal images of MØ in FITC-␣Bgtx-labeled cultures (Fig. 7B,  left panel). The receptor specificity of this staining is demonstrated by the disappearance of the fluorescent signal in cells preincubated with 1 M ␣Bgtx or 500 M nicotine (Fig. 7B,  right panel). Finally, MØ in culture were subjected to different treatments to evaluate if dup␣7 mRNA levels could be modulated. Fig. 7C shows that the pro-inflammatory cytokine interleukin-1␤ (IL-1␤), LPS from Salmonella abortus and nicotine, all markedly and significantly reduced dup␣7 mRNA expression in MØ cultures (Fig. 7C).

DISCUSSION
In this study we have cloned and expressed, for the first time, the full-length coding sequence of dup␣7 in GH4C1 cells and oocytes. Additionally, we have performed a functional study of the protein expressed in the latter cell type.
Our results indicate that dup␣7 acted as a dominant negative regulator of ␣7 nAChR activity through a mechanism involving reduction in the number of functional ␣7 nAChRs incorporated into the oocyte surface. Determining whether the dup␣7 subunit plays a functional role in vivo, wherever it is co-expressed with ␣7, requires additional experiments, but our preliminary results obtained in human cerebral cortex, human macrophages, and HL-60 cells suggest this possibility. Here, we discuss the most relevant results of this study.
The parallel confocal analysis of dup␣7 and ␣7 protein expression in transfected GH4C1 cells and in mRNA-injected oocytes labeled with the Mab306 antibody indicates that the two cell types express both foreign proteins, although their patterns and expression levels differ. Thus, GH4C1 cells show a similar high ␣7 and dup␣7 expression (Fig. 1, A and C), although a careful analysis of the ␣7 expression along the x axis of the cell reveals a peak localized in the cell membrane, although the dup␣7 expression peak seems to have an inner location, probably within the ER (Fig. 1B). Meanwhile, expression levels of dup␣7 on the cell surface of mRNA-injected oocytes represent only one-third of that of the ␣7 subunit (Fig. 2B), making it likely that most of the synthesized dup␣7 proteins in oocytes are retained in the ER. However, our data do not preclude the possibility that a small fraction of dup␣7 can form homo-or heteropentameric structures with other nAChR subunits able to migrate to the cell membrane.
Despite the low expression of dup␣7 compared with ␣7 in the oocyte surface, there is no doubt that the first protein clearly interferes with the second. This assertion is deduced from electrophysiological records of I Nic obtained in oocytes injected with different combinations of ␣7/dup␣7 mRNA (Fig. 3, A and B). Thus, although dup␣7 subunits are not capable of forming by themselves, homomeric nAChRs activated by ACh, or nicotine (not shown), they do produce a concentration-dependent dominant negative regulatory effect on ␣7 current. The question that now arises is to identify the mechanism through which dup␣7 exerts this effect. The following hypotheses, individually or in combination, could provide an answer. Thus, dup␣7 could do the following: 1) act at the transcriptional level of CHRNA7; 2) interfere with the proper oligomerization and assembly of ␣7 subunits in the ER, decreasing the number of mature ␣7 nAChRs able to migrate to the oocyte membrane; and/or 3) be integrated into a putative "␣7/dup␣7 heteromeric receptor" whose affinity or gating mechanism would be affected by the incorporation of the atypical subunit.
The first proposal can be ruled out, at least in terms of the dominant negative effect of dup␣7 on ␣7 nAChRs in oocytes because, in this case, the injected mRNAs are obtained by in vitro transcription from the corresponding cDNAs. Although our results cannot rule out the possibility that dup␣7 might produce some kind of interference in vivo at the transcriptional level of CHRNA7, this is unlikely because native transcripts that effectively regulate gene expression are "antisense" (44) and the orientation of the dup␣7 transcript in human brain is in the "sense" direction of transcription (26).
Collectively, our data tend to support the second hypothesis because dup␣7 co-expression significantly reduces the number of ␣7 nAChRs expressed on the oocyte surface, as can be inferred by the drastic decrease of fluorescent signal in Mab306-incubated oocytes as well as by the reduction of the B max in 125 I-␣ Bgtx binding assays (Fig. 4). This hypothesis is further reinforced by the similar degree of reduction of both I Nic (Ϸ 74%) and ␣7 nAChR expression (Ϸ70%) in FITC- ␣Bgtx-labeled oocytes injected with equimolar amount of ␣7 and dup␣7 mRNAs (Figs. 3B and 4B, right panel).
In relation to the third hypothesis, two of our findings seem to support this alternative in combination with the second proposal raised above. On the one hand, there was a small but statistically significant increase of K d in oocytes co-injected with the combination 1:4 of ␣7/dup␣7 mRNA compared with those injected with ␣7 mRNA alone (Fig. 4C, bottom panel). Because K d is the reciprocal of affinity, the higher the K d value, the lower the affinity of the receptor expressed by Bgtx. On the other hand, there was a different behavior of PNU120596 and 5HI in relation to their ability to enhance I ACh in oocytes injected with different combinations of ␣7/ dup␣7 mRNA (Fig. 5). Although the former drug distinguishes between different groups of oocytes, the second does not. These results can be interpreted in light of different ␣7 modulation sites required by one or the other positive allosteric modulators, because PNU120596 interacts with transmembrane domains whereas 5HI requires the extracellular N-terminal domain of the ␣7 subunit (40,45,46). Thus, PNU120596 could act on homo-and heteromeric receptors, whereas 5HI would do so only on homomeric ␣7 nAChRs.
Northern blot analysis shows a band of a size of 3.9 kb in human CTX and HL-60 cells that most likely corresponds to the dup␣7 transcript (Fig. 6A). This assumption is based on two facts as follows: 1) the HL-60 cells appear to express only dup␣7 transcript (25); and 2) analysis of the NCBI EST data base provides a theoretical prediction of the size of dup␣7 mRNA of 3600 bp. If we added the ϳ200 bp corresponding to the poly(A ϩ ) tail to this value, there is a good correlation between the predicted and the experimental size of dup␣7 mRNA found in this study. The possibility that the 3.9-kb band of CTX corresponds to an ␣7 isoform is very remote because we have not been able to find any isoform with this size in the literature or in the NCBI EST data base. In fact, the Northern blot for the CTX mRNA in our study shows additional bands that probably correspond to several ␣7 isoforms because their sizes are consistent with those found in SHY-5Y cells (47,48). Additionally, the GenBank TM sequence NM_000746 gives a size of the transcript encoded by the CHRNA7 gene of 3,351 bp, which could correspond to the band of 3.3 kb identified in our experiments. By identification of polysome-bound mRNAs, our results show, for the first time, that dup␣7 mRNA is being translated in HL-60 cells (Fig. 6B), a process that has not been demonstrated before because of the unavailability of a selective antibody.
After demonstrating that dup␣7 mRNA is translated and that the resulting protein behaves as a dominant negative of ␣7 nAChR activity in vitro, it is time to consider whether dup␣7 could play a relevant functional role in vivo in those tissues and cell types that also express ␣7 nAChRs, such as the central nervous system (49 -51) where presynaptic ␣7 nAChRs regulate the release of different neurotransmitters (1)(2)(3)(4)52). The same can be said of ␣7 nAChRs in MØ, which have an essential role in the cholinergic anti-inflammatory response (34,(53)(54)(55)(56). However, to assume that dup␣7 could behave as an endogenous modulator for ␣7 nAChR activity in vivo, it is essential that it meet two conditions as follows: (i) its mRNA levels must be high enough in respect to ␣7, and/or (ii) the expression levels of this messenger should, in turn, be susceptible to modulation by external stimuli that require greater or lesser ␣7 activity.
When we explored the first condition in human CTX and MØ using Q-PCR, we found that dup␣7 mRNA levels in the former tissue come to about one-fifth of those of ␣7 (Fig. 7A). This result is consistent with previous data obtained in the human prefrontal cortex showing that expression of ␣7 mRNA is five times higher than that of dup␣7 (21). The low proportion of dup␣7 relative to that of ␣7 mRNA in CTX is functionally relevant in accordance with our data from I Nic recorded in oocytes injected with the combination 5:1 of ␣7/ dup␣7 mRNA (Fig. 3, A and B).
A result opposite to that of CTX was obtained in MØ, which preferentially express dup␣7 over the ␣7 transcript (Fig. 7A). Human PBMC and lymphocytes (25,27,57) and HL-60 cells (Fig. 6A) behave similarly. Despite the low expression level of ␣7 mRNA in MØ, there is no doubt that it is efficiently translated given the high number of functional ␣7 nAChRs detected in our experiments (Fig. 7B) as well as in previous studies (34). Thus, the question now is why, despite the high level of dup␣7 mRNA in MØ, ␣7 nAChR expression is high in these cells. At this time we do not have a convincing answer, although different efficiencies in translation and/or proper folding of the ␣7 and dup␣7 proteins could be underlying these results. Additional experiments are needed to explore these possibilities.
In relation to the second condition for dup␣7 playing a modulatory role in vivo, our results in MØ show that mRNA levels are actually being regulated (Fig. 7C). This is an interesting finding because stimulation of ␣7 nAChR in macrophages induces a clear anti-inflammatory response secondary to decreased production of pro-inflammatory cytokines (34,53,56,58). In this context, our results show that one of these cytokines, IL-1␤, as well as LPS, is able to reduce dup␣7 mRNA expression by half. In these conditions, pro-inflammatory cytokines would repress dup␣7 expression and therefore increase the number of functional ␣7 nAChRs; the net output of this feedback loop would be the attenuation of the inflammatory response. Additionally, our findings that nicotine markedly reduces the expression levels of dup␣7 mRNA in MØ (Fig. 7C) are in complete agreement with recent in vivo results showing a lower ␣7 transcript expression in PBMC from smokers than from nonsmokers (57). Further experiments are needed to learn whether cerebral dup␣7 mRNA expression can be regulated by different stimuli, as occurs in MØ. However, it is interesting to note that the ratio of the ␣7/dup␣7 transcripts varies in different brain regions as well as in individuals addicted to drugs or suffering certain psychiatric disorders. This is the case in schizophrenic patients, who show a general decrease in brain ␣7 mRNA and all CHRNA7 transcription products that is most marked in the corpus callosum (26). The effect of nicotine in reducing dup␣7 mRNA levels found in our study could explain, at least in part, the relief of symptoms experienced by these patients after smoking. Also, in bipolar disorder, an imbalance in the ratio of CHRNA7/CHRFAM7A transcripts due to increased expression of dup␣7 mRNA has been found (21).
In summary, our data are consistent with the idea that the product resulting from the partial duplication of the CHRNA7 gene is likely to be modulated in a paracrine and endocrine manner, both in physiological as well as pathological situations, thereby exerting a negative modulating effect on ␣7 nAChR activity. Thus, previous findings regarding the CHRFAM7A and its possible association with psychiatric disorders can now be interpreted in the light of these new results.