Conservation within the RIC-3 gene family: Effectors of mammalian nicotinic acetylcholine receptor expression

Conservation within the RIC-3 Gene Family EFFECTORS OF MAMMALIAN NICOTINIC ACETYLCHOLINE RECEPTOR EXPRESSION* Received for publication, January 8, 2003, and in revised form, May 28, 2003 Published, JBC Papers in Press, June 23, 2003, DOI 10.1074/jbc.M300170200 Sarah Halevi‡, Lina Yassin‡, Margalit Eshel‡, Francisco Sala§, Salvador Sala§, Manuel Criado§, and Millet Treinin‡¶ From the ‡Department of Physiology, Hebrew University – Hadassah Medical School, Jerusalem 91120, Israel and the §Instituto de Neurociencias, Universidad Miguel Herna´ndez-CSIC, 03550-San Juan, Alicante, Spain In Caenorhabditis elegans, the ric-3 gene is required for the maturation of multiple nicotinic acetylcholine receptors (nAChRs), whereas other neurotransmittergated channels expressed within the same cells are unaffected by the presence of RIC-3. Here we show that RIC-3 is a member of a conserved gene family with representatives in both vertebrates and invertebrates. All members of this family have two transmembrane domains followed by a coiled-coil domain. Expression of the human ric-3 homolog, hric3, like the C. elegans ric-3, enhances C. elegans DEG-3/DES-2, rat ␣ 7, and human ␣ 7 nAChR-dependent whole-cell current amplitudes in Xenopus leavis oocytes, thus demonstrating functional conservation. However, hric3 also reduces human ␣ 4 ␤ 2 and ␣ 3 ␤ 4 nAChR-dependent whole-cell current amplitudes. Thus, hric3 shows differential effects on human nAChRs unlike the observed uniform effect of ric-3 on C. elegans nAChRs. Moreover, hric3 totally abolished currents evoked by 5-HT 3 serotonin receptors, whereas it barely modified ␣ 1 glycine receptor currents. With this caveat, RIC-3 belongs to a conserved family of genes likely to regulate nAChR-mediated transmission throughout evolution. Analysis of transcripts encoded by the hric3 locus shows that it encodes for multiple transcripts, likely to produce multiple hric3 isoforms, and that hric3 is expressed in neurons and muscles, thus enabling its interactions with nAChRs in vivo. Nicotinic acetylcholine receptors are widely expressed ligand-gated ion channels that mediate fast synaptic excitation and have additional roles including modulation of synaptic release (1). The nAChRs 1 are homomers or heteromers composed of five subunits. Each subunit traverses the membrane four times and is posttranslationally modified by both disulfide bond formation and glycosylation. Maturation of nAChRs, leading to production of fully assembled and functional receptors on the plasma membrane, is a complex, time-consuming, and poorly characterized process (2–4). Recently, we identified a Caenorhabditis elegans gene, ric-3, likely to be an important player in the maturation of nAChRs. In C. elegans, RIC-3 is required for cholinergic transmission mediated by nAChRs in neurons and in muscles but not for synaptic transmission mediated by other ligand-gated ion channels, even when these are expressed within the same cells as the nAChRs. RIC-3 activity is required within the cells that express the nAChRs and is likely to affect the processes of receptor folding or assembly within the endoplasmic reticulum (5). RIC-3 is a protein with two transmembrane domains followed by coiled-coil domains. When first identified, this protein showed no similarity to any characterized protein. Only one homolog was identified, the Drosophila CG9349 gene that is similar in both sequence and overall predicted structure. However, the demonstrated ability of RIC-3 to enhance whole-cell current amplitudes resulting from expression of both the C. elegans DEG-3/DES-2 and the rat ␣ 7 receptors in Xenopus oocytes suggested functional conservation of ric-3 in evolution (5). Here we show that ric-3 belongs to a conserved gene family with both vertebrate and invertebrate representatives. We proceeded to characterize the human ric-3 homolog (hric3) and showed that, like ric-3, it is capable of enhancing nAChR-dependent whole-cell current amplitudes in oocytes. Moreover, this effect corresponds with enhanced surface expression of the ␣ 7 receptor. Therefore, the ric-3 family is a conserved family of genes regulating nAChR maturation. However, hric3 had diverse effects on co-expressed nAChRs and even on another ligand-gated channel from the same ligand-gated ion channel superfamily. Specifically, although it enhances whole-cell current amplitudes produced by the C. elegans DEG-3/DES-2 nAChR and rat and human ␣ 7 nAChRs, it inhibits the wholecell current amplitudes produced by two other human nAChRs, ␣ 4 ␤ 2 and ␣ 3 ␤ 4, as well as a mouse 5-HT 3 serotonin receptor. Thus, hric3 shows a high diversity in its effects on co-expressed receptors. In addition, the hric3 locus encodes multiple transcripts coding for different protein isoforms. These transcripts are expressed widely in both excitable and non-excitable tissues. Such isoforms, if functional, may interact with hRIC3 to modulate its activity. EXPERIMENTAL PROCEDURES Sequence Analysis—ric-3 homologs were identified using the protein sequence of ric-3 as a probe in a standard tblastn search at www.ncbi. nlm.nih.gov/BLAST (6). Sequences for ric-3 homologs were obtained from: Danio rerio sequence CA475611, Ostertagia ostertagi sequence BG734142, Drosophila melanogatser CG9349 encoded by FBg0034574, a reconstruction of two Xenopus leavis sequences, AW645559 and AW634327, the human clone AL520122 (clone FL1001, Invitrogen), which was sequenced in the Life Science Sequencing Facilities (Hebrew University) using internal primers (sequencing was done in both orien- * This work was supported by a U. S.-Israel Binational Science Foundation Grant 1999-074-01 and grants from the Ministry of Education of Spain (Grants PM98-0097 and PM98-0104) and Generalitat Valenciana (Grant CTIDIB/2002/138). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to GenBank™/EBI Data Bank with the accession number(s) AY326435 and AY326436. ¶To whom correspondence should be addressed. Fax: 972-2-6439736; E-mail: [email protected]. 1 The abbreviations used are: Ach, acetylcholine; nAChR, nicotinic Ach receptor; 5-HT 3 , serotonin (5-hydroxytryptamine) 3; 5-HT 3 R, 5-HT 3 receptor; EST, expressed sequence tag; Bgt, bungarotoxin; GluR3, glutamate receptor 3; GlyR, glycine receptor; h, human. THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 36, Issue of September 5, pp. 34411–34417, 2003 © 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. This paper is available on line at http://www.jbc.org 34411 This is an Open Access article under the CC BY license. tations (accession number AY326435)), and the mouse expressed sequence tags (EST) sequence BB642650 (the mouse sequence is a reconstruction of this EST sequence and the corresponding genomic sequence, a reconstruction made possible due to the strong conservation of hric3 and mouse ric3 nucleotide sequences). Sequence alignment was generated using Clustal_W analysis (McVector). Structure prediction analysis was done using programs on www.expasy.ch and psort.nibb.ac.jp. Analysis of alternative hric3 transcripts and the resulting protein isoforms is based on sequences available in the databases, except clone BF680796 (IMAGE:4296897), which was sequenced in the Life Science Sequencing Facilities (Hebrew University) using internal primers (accession number AY326436). This cDNA is likely to be incomplete as it does not have a poly(A) sequence at the 3⬘-end. When a single base mismatch was found between different available versions of the same sequence, we chose the version that maintained continuity of the open reading frame. Heterologous Expression and Electrophysiology—A 1.2-kb XhoIEcoRV fragment containing hric3 from AL520122 was cloned into pGEMH19, which contains Xenopus globin untranslated sequences. DEG-3, DES-2 (7), and GluR3(L507Y) (8) clones were described previously (5). DNAs of human neuronal nAChR subunits ( ␣ 7, ␣ 3, ␣ 4, ␤ 2, and ␤ 4) (9), human GlyR ␣ 1 (10) subunit, and mouse 5-HT 3 R (11) were inserted into the pSP64T vector (12). In vitro transcribed and capped cRNAs were injected at final concentrations of 0.1–0.05 ng for DEG-3 and DES-2, 0.2 ng for GluR3(L507Y), and 5 ng for the mammalian nAChR subunits, except ␣ 7, GlyR ␣ 1, and 5-HT 3 R, which were used at 10 ng. Subunits of heteromeric receptors (DEG-3/DES-2, ␣ 4 ␤ 2 and ␣ 3 ␤ 4) were injected in an equimolar ratio. Optimal effects of hric3 on DEG-3/DES-2 expression were found at 1:5 hric3 to DEG-3/DES-2 concentrations, but for mammalian nAChR experiments, equal concentrations (5 ng in all cases) were injected. The final concentrations of subunits cRNAs were kept constant between injections with or without hric3 for each experiment. Injections and recordings were done as described in Treinin et al. (13). The effects of hric3 co-expression on receptor-mediated whole-cell current amplitudes were assayed between the first and third day following injections. Errors are standard error of means, and significance was examined using the paired ttest or the Mann-Whitney test. Specific surface expression of [ 125 I] ␣ -Bgt binding sites was tested with 5 nM[ 125 I] ␣ -Bgt as described (14). Briefly, oocytes were incubated with 5 nM[ 125 I] ␣ -Bgt for1hatroom temperature. At the end of the incubation, unbound [ 125 I] ␣ -Bgt was removed, oocytes were washed, and bound radioactivity was counted. Non-specific binding was determined using non-inoculated oocytes. Northern and in Situ Analysis—Expression of hric3 was examined on human multitissue Northern blot number 7760-1 (Clontech), using random primed labeled cDNA probes for the entire hric3 open reading frame or the first two exons. For in situ analysis, we used 35 S-labeled in vitro transcribed cRNA from a SKII ⫺ clone containing a 0.6-kb mouse ric3 fragment. This EcoRV-BamHI fragment is derived from clone BU705427 (IMAGE:6408852). Both the sense (T7) and the antisense (T3) cRNAs were used as probes on wax sections of mouse brain (15). This analysis showed that the antisense probe produced a stronger and less uniform signal. RESULTS A Conserved Family of ric-3-like Proteins—Blast searches of the available databases of ESTs have identified ric-3 homologs in a number of species, including the invertebrates O. ostertagi (a parasitic nematode) and D. melanogaster, and the vertebrates X. leavis, D. rerio (Zebrafish), Mus musculus, and Homo sapiens (Fig. 1). All sequences show high homology in the two membrane-spanning domains and in the first coiled-coil domain (Fig. 1). In addition, all members of the family have a proline-rich region (⬎10% prolines) in between the two transmembrane domains. For all homologs, structure prediction analysis suggest that both the N terminus and the C terminus coiled-coil domain are cytosolic. Thus, they all conform to the suggested structure and topology of the C. elegans ric-3, which would allow the formation of a coiled-coil-mediated protein complex in the cytosol (5). Although the basic structure is conserved within the RIC-3 family, some differences exist. Only the C. elegans ric-3 has a long N-terminal domain and a second coiled-coil domain; however, since the sequences of O. ostertagi, D. rerio, and X. leavis are incomplete, we cannot exclude the possibility that they also code for a second coiled-coil domain (or specifically for O. ostertagi, a long N-terminal domain). Differences are also found in the Drosophila ric-3, which has longer spacer sequences between the two transmembrane domains and between the second transmembrane domain and the coiled-coil domain (Fig. 1B). A Human Homolog of ric-3—To further characterize this gene family, we focused on the H. sapiens ric-3 homolog (hric3). For this purpose, we obtained a full sequence of one of the available clones, AY326435, a clone obtained from a neuroblastoma cell line. This cDNA is 2.9-kb-long, coding for a 369amino-acid-long open reading frame and a long 3⬘-untranslated region. The corresponding genomic sequence spans 60 kb on chromosome 11p15,4 and is encoded by 6 exons. A similar sequence (85% of amino acids are identical) with a similar genomic structure is found in M. musculus on chromosome 7. The full sequence of hric3 conforms to the general structure prediction for the ric-3 family having two transmembrane domains followed by a coiled-coil domain (Fig. 1B). Specifically, overall identity with the C. elegans ric-3 is 22%, and overall structure differs by having a shorter N-terminal domain and only one coiled-coil domain instead of the two found in ric-3 (5). Functional Conservation of hric3—Human ric3 shows significant similarity to C. elegans ric-3 in both sequence and predicted structure. However, such sequence conservation does not imply functional conservation. To assay for functional conservation, we examined the effects of co-expression with hric3 on the whole-cell current amplitudes produced by ion channels expressed in Xenopus oocytes. Injection of non-saturating concentrations of the C. elegans DEG-3 and DES-2 cRNAs leads to reproducible choline-dependent whole-cell currents (131 ⫾25 nA, n⫽28 oocytes, N⫽6 frogs). Co-expression with hric3 enhances these currents 5-fold (631 ⫾86 nA, n⫽28, N⫽6). Co-expression of hric3 with GluR3 does not increase the amplitude of glutamate-activated whole-cell currents, nor does expression of hric3 alone lead to detectable choline, nicotine, or glutamate-activated currents (Fig. 2). Thus, hRIC3, like the C. elegans RIC-3, appears to specifically enhance nAChR-dependent whole-cell current amplitudes. We have previously shown that C. elegans RIC-3 enhances whole-cell current amplitudes elicited by the rat ␣ 7 neuronal nAChR. AY326435 was found in a neuroblastoma cell line, suggesting expression of hric3 in neurons (further evidence for co-localization of ric3 with ␣ 7 is provided below). Thus, hric3 is likely to be expressed in the same cells as ␣ 7, enabling interactions between the two proteins in vivo. Indeed co-expression of hric3 with rat (not shown) and human ␣ 7 leads to a 2-fold increase in acetylcholine (ACh)-induced whole-cell current amplitudes (Fig. 3C). Moreover, direct measurement of ␣ 7 nAChR expression on oocyte membranes as detected by [ 125 I] ␣ -Bgt binding indicated an increase of about 6-fold in binding sites in the presence of hric3 (Fig. 3B), supporting its role in ␣ 7-maturation in vivo. The hric3-dependent enhancement in current amplitude is not associated with changes in current kinetics (Fig. 3A). Mammals utilize additional types of nAChR that could also be targets for hRIC3 activity. Thus, it was interesting to examine the effects of hric3 expression on two abundantly expressed neuronal nAChRs. Co-expression of hric3 with both ␣ 4 ␤ 2 and ␣ 3 ␤ 4 produced marked inhibition of activity, down to 1.5 and 55%, respectively, as compared with controls without hric3 (Fig. 3C). Current kinetics observed with ␣ 3 ␤ 4 nAChRs were not altered by the presence of hric3 (Fig. 3A), and ACh doseresponse curves were also similar (not shown), indicating that inhibition was not due to a decreased affinity for ACh. These results suggest inhibitory interaction of hRIC3 with these two hric3 Regulates Nicotinic Receptor Expression34412 heteromeric nAChRs. Such inhibition was not seen in C. elegans, where all four nAChRs examined required RIC-3 for their activity (5). It thus appears that humans differ from C. elegans in having a more specialized role for hRIC3 in cholinergic transmission. In fact, these differences extend to the effect of hRIC3 on other ligand-gated receptors. Thus, co-expression of hRIC-3 with FIG.1.The ric-3 gene family. A, sequence alignment of RIC3 homologs. Underlined are the two transmembrane domains (single line) and the coiled-coil domain (double line). Sequences in the N terminus of the C. elegans, O. ostertagi, and Drosophila genes (33, ⬎3, and 14 amino acids, respectively) and the C termini of all homologs are not shown as these regions show no significant homology. When all sequences are aligned together, the coiled-coil region of the Drosophila gene does not align. A separate alignment done to the coiled-coil domains of the C. elegans and Drosophila genes does, however, show significant identity (see panel B). B, structure and conservation within the ric-3 gene family. Numbers indicate percentage of identity for each domain in the RIC3 homologs relative to the C. elegans ric-3.Boxes indicate the different domains. Transmembrane domains are indicated with TM1 and TM2. Coiled-coil domains are indicated with CC1 and CC2; CC1 alone is conserved. The spacer domain between the two transmembrane domains, the proline-rich spacer (PRS), has ⬎10% prolines in its sequence. The O. ostertagi, D. rerio (Zebrafish), and X. leavis (Xenopus) sequences are incomplete. Thus, we have no information on the length and structure of their C termini, neither do we know the length and structure of the O. ostertagi N termini. hric3 Regulates Nicotinic Receptor Expression 34413 5-HT 3 receptors produced almost total inhibition of currents elicited by serotonin (Fig. 3, Dand E). By contrast, another homomeric receptor, the ␣ 1 GlyR, showed no differences in glycine-dependent whole-cell currents (Fig. 3, Dand E). Complex Expression Pattern of the hric3 Locus—To validate expression of hric3 in brain, we used a commercially available multitissue Northern blot (Fig. 4). AY326435, a full-length cDNA coding for the hric3 homolog, is 2.9 kb in length. A 394-bp probe from the first two exons of hric3 identifies such a transcript in excitable tissues, muscle, brain, and heart. However, this same probe also identifies additional transcripts that are expressed in non-excitable tissues. To understand the nature of these transcripts, we analyzed EST and mRNA sequences available in the databases. Indeed, the human genome project identified multiple mRNA sequences encoded by this locus. (a) Sequence BC022455 (hypothetical protein FLJ1608) is identical to the sequence of AY326435 except for the absence of3bpatthe5⬘-end of exon 4 (leading to the absence of a single serine at this site), a difference that may be attributed to editing and alternative splicing (Fig. 5A) (16). The length of the transcript encoding for this alternative isoform is the same as AY326435. Thus, it is not clear which of these transcripts contributes to the 2.9-kb band seen in brain, heart, and skeletal muscle. (b) Transcript AK021670 is 1.5 kb encoding for a soluble coiled-coil domain (Fig. 5B). The expression pattern of this isoform is presently unclear since probing with the entire hric3 open reading frame (1.2-kb probe, not shown) or the first two exons alone (Figs. 4 and 5B, the AK021670 transcript starts in the second intron) revealed the same transcripts. Thus, this transcript may be rare or may co-migrate with another transcript. (c) Transcript AL832601 is 5.2 kb encoding for the first membrane-spanning domain spliced directly to the C terminus (Fig. 5C). A transcript of the same size as AL832601 is strongly expressed in the pancreas. (d) The transcript corresponding to ESTs AY326436 (prostate) and BI832705 (pancreas and spleen) is a product of read-through into the second intron, thus encoding a truncated protein coding for the two trans-membrane domains only. Length of this transcript is not known since the available cDNA appears to be incomplete (Fig. 5D). The complex expression pattern of the hric3 locus is not unusual as others have shown that mammalian loci encode for multiple transcripts and have suggested that many of these FIG.2. hRIC3 co-expression enhances nAChR-dependent whole-cell peak current amplitudes in Xenopus oocytes. The effects of hric3 expression on peak whole-cell-current amplitudes elicited by 3.2 mMcholine on oocytes expressing DEG-3/DES-2 channel,1 mMglutamate on oocytes expressing GluR3 (L507Y) channel, or 3.2 mM choline on oocytes expressing hRIC3 alone were measured in voltage clamped oocytes (holding potential ⫺70 mV). A, representative traces. Solid bars indicate time of agonist application. Notice the differences in amplitude scales. B, cumulative responses with co-expression of hRIC3. DEG-3/DES-2, n⫽28, N⫽6; GluR3, n⫽30, N⫽5; and hRIC3 alone, n⫽20, N⫽4. *** represents statistical significance (p⬍0.001) calculated by the paired ttest. FIG.3.Differential effects of hRIC3 co-expression on mammalian receptors. A, the effects of hric3 co-expression on whole-cell currents elicited by the continuous application of 1 mMACh in oocytes expressing human nAChRs ( ␣ 7, ␣ 4 ␤ 2, and ␣ 3 ␤ 4). Traces are representative currents obtained at a holding potential of ⫺80 mV. The vertical scale bar is the same for all types. Notice that the time scale for ␣ 7-expressing oocytes is shorter. B, the effect of hric3 co-expression on ␣ -bungarotoxin binding sites in oocytes expressing homomeric ␣ 7 nAChRs. C, summary of peak current amplitudes obtained under the conditions of panel A. Individual peak amplitudes (both control and hric3-co-expressing oocytes) were normalized to the arithmetic mean of the peak amplitudes obtained in control oocytes. Data represent mean and standard error of 15–53 oocytes (N⫽3–6 frogs). D, the effects of hric3 co-expression on whole-cell currents elicited by the continuous application of either 100 ␮ Mserotonin or 1 mMglycine in oocytes expressing homomeric mouse 5-HT 3 receptors or human ␣ 1 glycine receptors, respectively. Holding potential was ⫺80 mV. E, summary of peak current amplitudes obtained under the conditions of panel D. Data were normalized as in panel C. Data represent mean and standard error of 10–20 oocytes (N⫽2 frogs). In B,C, and E, *** represents the statistical significance (p⬍0.001) calculated by the Mann-Whitney test. hric3 Regulates Nicotinic Receptor Expression34414 transcripts are non-functional (17). However, it is interesting to speculate on the roles of these isoforms, if functional, since their differential expression could produce complex regulatory interactions. For example, the soluble coiled-coil domain may compete with the membrane-bound coiled-coil for interacting proteins, thereby inhibiting its activity. Localization of ric3 Transcripts in Mouse Brain—In humans, expression of hric3 transcripts is seen in many tissues, and hric3 has diverse effects on ligand-gated ion channel expression. Thus, detailed analysis of the distribution of ric3 is important to identify its in vivo targets. For this purpose, we used a mouse probe, encompassing the second membrane domain and the coiled-coil domain, for in situ analysis on mouse brain. Using this probe, we could detect a low signal from most regions of the brain and a stronger signal from few isolated regions: the CA1–CA3 region in the hippocampus (Fig. 6, Aand B), the deep nuclei and the Purkinje cell layer in the cerebellum (Fig. 6, Cand D), and the superior colliculus (not shown). These three regions overlap with regions of high ␣ 7 expression (18), enabling interaction in vivo between the two proteins. However, it is interesting to note that not all ␣ 7-expressing regions express high levels of ric3, as seen for the dentate gyrus (Fig. 6, Aand B). DISCUSSION Although we first identified ric-3 as coding for a novel protein with no similarity to any other characterized proteins in the databases (5), here we show that ric-3 actually belongs to an evolutionarily conserved gene family. This conservation is also functional since the human homolog hric3, like the C. elegans ric-3, enhances whole-cell current amplitudes for some nAChR. In C. elegans ric-3 is required for cholinergic transmission mediated by the neuronal DEG-3/DES-2, the body muscle levamisole-sensitive, the body muscle nicotine-sensitive, and the pharyngeal-muscle EAT-2 nAChRs. These receptors differ in their subunit composition, function, and pharmacology. Moreover, RIC-3 is capable of enhancing peak current amplitudes elicited by the rat ␣ 7 nAChR in Xenopus oocytes. This apparent lack of discrimination between different nAChRs suggests that RIC-3-mediated effects depend on nAChR-specific domains, common to different receptor subtypes, and therefore likely to be conserved in evolution (5). Similarly, the human counterpart of ric-3,hric3, enhances the peak whole-cell current amplitudes elicited by both the C. elegans DEG-3/DES-2 receptor and the human and rat ␣ 7 receptors. However, co-expression with the human ␣ 4 ␤ 2 and ␣ 3 ␤ 4 receptors leads to reduction in wholecell current amplitudes, suggesting that the interaction between ric-3 homologs and members of the nAChR family was modified during evolution. Similar effects are also observed when the C. elegans ric-3 is used instead of hric3 (data not shown), which may indicate that potential structural variations within the mammalian nAChR family are responsible for the different behavior of the ␣ 3 ␤ 4 and the ␣ 4 ␤ 2 nAChR subtypes. Interestingly, the actions of hric3 appear to extend to other members of the same ligand-gated receptor gene superfamily. In this case, dramatic differences are also observed since expression of the closely related 5-HT 3 receptors is totally FIG.4.Expression pattern of hric3 transcripts. Human multitissue Northern blot probed with a 350-bp probe corresponding to the first two exons of the AY326435 transcript. The arrow indicates transcripts of the same size as AY326435. FIG.5. Multiple isoforms encoded by the hric3 locus. A, isoforms encoded by clones BC022455 and AY326435, both 2.9 kb in length with one difference: the absence in BC022455, or the presence in AY326435, of a single serine that is boxed.B, an isoform encoded by clone AK021670, 1.5 kb in length, coding for the coiled-coil domain only. C, an isoform encoded by clone AL832601, 5.2 kb in length, coding for the first transmembrane domain spliced directly to the N terminus. D, an isoform encoded by clone AY326436, coding for a two-transmembrane domain-only protein. cDNAs encoding for this transcript are incomplete; thus, the length of the transcript encoding for this isoform is unknown. In all cases, lines represent transmembrane domains, dotted lines represent coiled-coil domains, arrowheads represent the site of intron/exon boundary, and double lines represent sequences that are not found in AY326435. hric3 Regulates Nicotinic Receptor Expression 34415 inhibited by hric3, whereas no modification in expression of the more distantly related ␣ 1 glycine receptor is observed. 2 These differential effects could be explained in several ways. One possibility is that formation of a complex with hric3 (mediated by a conserved domain within all the affected receptor subunits) leads to arrest of maturation if it is not followed by additional interactions of the receptor subunits with hric3. These additional interactions, produced with ␣ 7 nAChRs but not with ␣ 3 ␤ 4or ␣ 4 ␤ 2 nAChRs, as well as with 5-HT 3 receptors, are likely to be mediated by non-conserved domains within the former subunits. Another possibility suggests that formation of “productive complexes”depends, in addition to a conserved interaction of hric3 with receptor subunits, on yet unidentified proteins that interact differentially with mammalian members of this gene superfamily. Both possibilities suggest that interaction of “non-compatible”receptor subunits with hric3 leads to a “dead-end”complex incapable of further maturation and adds a potentially new mechanism for regulating ligand-gated receptor expression. Here it is important to note that glutamate receptors, which are members of a different ligand-gated receptor superfamily, and ␣ 1 GlyR, a distant relative of the nAChRs, are not affected by hric3. These last results suggest specificity of the effects of hric3 on nAChRs and closely related receptors, a specificity that is similar to what is seen in C. elegans. The RIC-3 family is a conserved gene family effecting nAChRs maturation. The exact role of these proteins in nAChR maturation is yet unclear. In C. elegans, we showed that RIC-3 is found inside the cells, probably on endoplasmic reticulum membranes, and not at synapses. We also showed that in its absence, transport of the DEG-3 protein to the processes is greatly reduced (5), a result consistent with the effects that we now show on ␣ 7 surface expression. These results are consistent with a role for RIC-3 in folding or assembly, processes occurring in the endoplasmic reticulum, or in trafficking from the endoplasmic reticulum. We now show that ric-3 belongs to a conserved gene family. All members of this family are relatively small proteins (350–400 amino acids), showing no homology to enzymes likely to modify nAChRs. Conservation within this family is limited to having two transmembrane domains, separated by a proline-rich spacer and followed by a coiled-coil domain. Structure predictions suggest that both the N terminus and the C terminus coiled-coil are cytosolic. Based on this analysis, we suggest a role for RIC-3 proteins as adaptor proteins, bringing together receptors and proteins needed for their maturation. We also suggest that multiple domains within the RIC-3 homologs mediate interactions with components of such a “maturation complex.”Coiled-coil domains have been implicated in protein-protein interactions (20). The prolines within the spacer between the two transmembrane domains, some of which appear at conserved positions, may provide a structural scaffold on which to anchor interacting proteins. The membrane-spanning domains may also mediate protein-protein interactions. Here it is interesting to note that the second transmembrane domain in RIC-3 homologs shows conservation beyond what is needed for traversing the membrane. For example, within this domain are found a proline and a glycine that show 100% conservation. Also, the membranespanning domains of ligand-gated receptors are relatively well conserved and may form the conserved recognition domain for RIC-3 homologs. Thus, interactions within the membrane may bring together receptor subunits and the RIC-3 homologs, whereas other RIC-3 domains may be needed for additional interactions (the non-conserved interaction suggested above) and for interactions with yet unidentified proteins, thus forming a multiprotein maturation complex. Analysis of transcripts encoded by the hric3 locus shows multiple transcripts resulting from editing, alternative splicing, and alternative promoters. These transcripts are differentially expressed in many tissues, including non-excitable ones. At this stage, the relevance of this complexity is unclear. Three options are possible. First, these additional transcripts produce functional proteins that differentially affect the machinery that matures nAChRs. Second, these transcripts produce functional proteins whose function is unrelated to nAChR maturation. Third, these shorter isoforms are non-functional, as suggested by Mallya et al. (17). It is important to note that expression of these transcripts in non-excitable tissues and in heart muscle is not sufficient evidence for roles other than regulation of nAChRs, as expression and function of nAChRs was shown in many tissues including non-excitable tissues (21). In C. elegans,ric-3 is a central player in cholinergic transmission mediated by nAChRs. In its absence, cholinergic transmission in both muscles and neurons is severely compromised. Here we show that in mammals, the role of hric3 is likely to be more general and heterogeneous. However, the complex expression of the hric3 locus, as well as the lack of knowledge on functional expression of hric3 and its isoforms, preclude conclusions as to the role of hric3 in nAChR-mediated transmission in humans. With these reservations in mind, a likely interpretation of our data is that hric3 regulates expression of ␣ 7 receptors, which are known to depend on host-specific factors for their expression (19, 22). Indeed, mouse brain regions expressing high levels of ric3 transcripts are known to express ␣ 7, thus enabling an interaction between the two proteins (18). It is also possible to speculate that the differential interactions of hric3 with human ligand-gated ion channels are exploited in vivo to differentially regulate expression of these receptors. Acknowlegements—The help of L. M. Valor and J. Mulet in the mammalian nAChRs experiments, of H. Farah in the C. elegans nAChR experiments, of Drs. A. Klar and K. Marom in the in situ analysis, and of Dr. E. Keshet in photography is acknowledged. We thank Drs. J. Lindstrom, M. Koenen, and H. Betz for the human nAChRs, mouse 5-HT 3 receptor, and human ␣ 1 GlyR cDNAs, respectively. 2 For more information, see www.pasteur.fr/recherche/banques/ LGIC/catphylogen.html. FIG.6.ric3-expressing regions in mouse brain. In situ analysis of ric3 expression in mouse brain using a 0.6-kb mouse ric3 probe encompassing the second transmembrane domain and the coiled-coil domain. A, a bright field image of the hippocampus. B, a dark field image of the hippocampus showing strong ric3 expression in CA1–CA3 but not in the dentate gyrus (DG). C, a bright field image of the cerebellum. D, a dark field image of the cerebellum showing strong ric3 expression in the Purkinje cell layer (PCL) and the deep nuclei (DN). hric3 Regulates Nicotinic Receptor Expression34416 REFERENCES 1. Clementi, F., Fornasari, D., and Gotti, C. (2000) Eur. J. Pharmacol. 393, 3–10 2. Merlie, J. P., and Lindstrom, J. (1983) Cell 34, 747–757 3. Green, W. N. (1999) J. Gen. Physiol. 113, 163–169 4. Keller, S. H., and Taylor, P. (1999) J. Gen. Physiol. 113, 171–176 5. Halevi, S., McKay, J., Palfreyman, M., Yassin, L., Eshel, M., Jorgensen, E., and Treinin, M. (2002) EMBO J. 21, 1012–1020 6. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J. Mol. Biol. 215, 403–410 7. Yassin, L., Gillo, B., Kahan, T., Halevi, S., Eshel, M., and Treinin, M. (2001) Mol. Cell. Neurosci. 17, 589–599 8. Stern-Bach, Y., Russo, S., Neuman, M., and Rosenmund, C. (1998) Neuron 21, 907–918 9. Nelson, M., and Lindstrom, J. (1999) J. Physiol. (Lond.)516, 657–678 10. Grenningloh, G., Schmieden, V., Schofield, P. R., Seeburg, P. H., Siddique, T., Mohandas, T. K., Becker, C. M., and Betz, H. (1990) EMBO J. 9, 771–776 11. Uetz, P., Abdelatty, F., Villarroel, A., Rappold, G., Weiss, B., and Koenen, M. (1994) FEBS Lett. 339, 302–306 12. Krieg, P. A., and Melton, D. A. (1984) Nucleic Acids Res. 12, 7057–7070 13. Treinin, M., Gillo, B., Liebman, L., and Chalfie, M. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 15492–15495 14. Garcı´a-Guzma´n, M., Sala, F., Sala, S., Campos-Caro, A., and Criado, M. (1994) Biochemistry 33, 15198–15203 15. Wilkinson, D. G., Bailes, J. A., Champion, J. E., and McMahon, A. P., (1987) Development 99, 493–500 16. Grauso, M., Reenan, R. A., Culetto, E., and Sattelle, D. B. (2002) Genetics 160, 1519–1533 17. Mallya, M., Campbell, R. D., and Aguado, B. (2002) Genomics 80, 113–123 18. Dominguez del toro, E., Juis, J. M., Peng, X., Lindstrom, J., and Criado, M. (1994) J. Comp. Neurol. 349, 325–342 19. Sweileh, W., Wenberg, K., Xue, J., Forsayeth, J., Hardy, S., and Loring, R. H. (2000) Mol. Brain Res. 75, 293–302 20. Stefancsik, R., Jha, P. K., and Sarkar, S. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 957–962 21. Sharma, G., and Vijayaraghavan, S. (2002) J. Neurobiol. 53, 524–534 22. Cooper, S. T., and Millar, N. S. (1997) J. Neurochem. 68, 2140–2151 hric3 Regulates Nicotinic Receptor Expression 34417