|Molecular Vision 2004;
Received 1 June 2004 | Accepted 18 October 2004 | Published 29 October 2004
Tissue specific expression of alternative splice forms of human cyclic nucleotide gated channel subunit CNGA3
Steven C. Cassar,
Neuroscience Research, Global Pharmaceutical Research and Development, Abbott Laboratories, Abbott Park, IL
Correspondence to: Steven C. Cassar, Neuroscience Research, Department 47W, Building AP9A-1, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL, 60064-6117; Phone: (847) 937-9302; FAX: (847) 937-9195; email: email@example.com
Purpose: The cyclic nucleotide gated cation channel subunit, CNGA3, (also known as the cone alpha CNG subunit), plays a vital role in signal transduction of human cone cells. Owing to its well established role in vision, the human CNGA3 isoform studied thus far was cloned from retinal tissue. However, non-human homologs of CNGA3 have been cloned from a variety of other tissues including kidney, heart, pineal gland, adrenal gland and testes. Among these, alternative splice forms of CNGA3 have been identified. The objective of this study was to explore alternative splice forms of human CNGA3 and determine their distribution among various human tissues.
Methods: RT-PCR was used to amplify full length open reading frames of CNGA3 from human testes RNA and to detect and distinguish among splice forms in 23 tissues. DNA sequencing was used to characterize full length splice forms and to confirm the identity of RT-PCR products from a number of tissues.
Results: Two new full length alternatively spliced forms of hCNGA3 (referred to as Variant 2 and 3) were isolated. These splice variants are different from the cone hCNGA3 (Variant 1) in that they lack exon 5. In addition, they differ from each other in that Variant 3 contains an extra exon that originates from the intron preceding exon 4. We demonstrate that CNGA3 transcripts are detectable in all 23 human tissues examined. In all tissues, except retina, Variant 2 specific PCR products had the brightest band intensity. In retina, the band from Variant 1 (containing exon 5) was most intense. In fact, among all tissues examined, Variant 1 can only be detected strongly in the retina.
Conclusions: The extensive distribution of CNGA3 and the tissue specific expression of alternative splice forms indicate widespread and diverse roles for CNGA3. The unique expression of Variant 1 in the retina implies a significance to the amino acids encoded by exon 5 that may be necessary for the function of CNGA3 in human cone cells.
Cyclic nucleotide gated (CNG) channels are ligand gated ion channels that open and close in response to changes in the intracellular concentration of the second messengers, 3,5-cyclic adenosine monophosphate (cAMP) and 3,5-cyclic guanosine monophosphate (cGMP). CNG channels play fundamental roles in signal transduction of photosensory and chemosensory cells and serve as key players in the function of rods, cones and olfactory sensory neurons. Native CNG channels are formed through complexes of alpha and beta subunits. When expressed alone in heterologous systems, the alpha subunits form functional channels whereas the beta subunits have been shown to modulate the properties of homomeric alpha subunit channels .
In photoreceptors, CNG channels elicit a membrane potential change in response to chemical signals produced by the absorption of light. Rod and cone cells express different CNG channels. Mutations in genes encoding the alpha (CNGA1) and beta (CNGB1) subunits of the rod CNG channel have been associated with autosomal recessive retinitis pigmentosa, characterized by photoreceptor degeneration and various forms of blindness [2-4]. Transcripts for rod CNG channel subunits have been detected in tissues other than retina in a variety of organisms [5-8]. Mutations in genes encoding the alpha (CNGA3) and beta (CNGB3) subunits of the cone CNG channel have been associated with autosomal recessive achromatopsia, characterized by perturbations in light sensitivity and total color blindness [9-12]. Transcripts for the cone CNG channel subunits have also been detected in non-retinal tissues in a variety of organisms. However, due to its essential role in vision, studies on CNGA3 in humans have, up to now, focused only on its expression in cone cells.
CNGA3 was originally cloned from chicken retina , and its expression has been detected in heart, kidney, brain, olfactory sensory neurons, pineal and adrenal gland from various species [5,14-17]. It has also been cloned from bovine sperm  where it is expressed on the sperm flagellum . CNG channels expressed on the flagellum of sperm have been shown to conduct calcium, suggesting a role in sperm motility and function . In this study, we report on the identification of two splice variants of human CNGA3 isolated from testes RNA and on the evaluation of tissue expression of these variants along with the originally reported form [20,21] in diverse human tissues.
CNGA3 cDNA was amplified by RT-PCR from human testes RNA (BD Biosciences, Palo Alto, CA). Oligo dT12-18 primer was used in conjunction with Superscript II Reverse Transcriptase (Invitrogen, Carlsbad, CA) and 5 μg total RNA to generate single stranded cDNA used as template in PCR. PCR was carried out using Elongase (Invitrogen). Oligonucleotide primers were designed based on the human sequence from cone (GenBank Accession number AF065314). The forward primer, 5'-TTT GGA TCC GCC GCC ATG GCC AAG ATC AAC ACC CAA TAC TCC CA-3' and reverse primer, 5'-TTT GGA TCC ACA TAC GGT CAC CCT GAC AGT CGA CCC T-3', flank the coding sequence and provide BamHI cloning sites. Thermocycling was done in a Stratagene Robocycler (La Jolla, CA) with the following parameters: an initial denaturation of 96 °C 30 s, 40 cycles of (96 °C 30 s, 60 °C 1 min, 72 °C 3 min), a final extension of 72 °C 7 min. RT-PCR products were visualized on an 0.8% agarose gel stained with ethidium bromide. Clones were obtained directly from the RT-PCR products in pcDNA3.1/V5-His-TOPO (Invitrogen) and were examined by sequencing.
Analysis of hCNGA3 expression was carried out with 23 human tissues by RT-PCR. Generation of cDNA from total RNA used random hexamers in conjunction with Superscript II Reverse Transcriptase (Invitrogen). Total RNA from dorsal root ganglia was obtained from Advanced Biological Services (Wilmington, Delaware). All other RNAs were from BD Biosciences. Retina and testes cDNA were purchased pre-prepared from poly-A selected RNA in the form of Quickclone cDNA (BD Biosciences). Testes cDNA was also prepared from total RNA (BD Biosciences) for comparison to the Quickclone sample. A negative control RT-PCR, identical in all aspects to the sample reaction, except lacking reverse transcriptase, was set up for each tissue. Oligonucleotide primers were designed to detect and distinguish splice variants of hCNGA3. PCR was performed using Platinum Taq DNA Polymerase High Fidelity (Invitrogen). Thermocycling was done in a Stratagene Robocycler with the following parameters: an initial denaturation of 96 °C 30 s, 40 cycles of (96 °C 30 s, 55 °C 1 min, 72 °C 30 s). The primers listed in Table 1 were used; their names indicate the region of the cDNA to which they anneal and the direction of their extension by polymerase. These primers anneal to regions of the cDNA that flank sequences between exons 3 and 4 (set A) or between exons 4 and 7 (set B; Figure 1). Thus, they can be used to detect the presence of variants that differ in their exonic structure in the targeted regions. Products were visualized on ethidium bromide stained 4% NuSieve 3:1 (BioWhittaker Molecular Applications, Rockland, MD) agarose gels.
Two full length isoforms, termed Variant 2 and Variant 3, of the hCNGA3 transcript were isolated from testes RNA. These forms differ from the hCNGA3 of the cone photoreceptors [20,21] (herein referred to as Variant 1) in that they both lack the sequence encoded by exon 5. In addition, they differ from each other in that one of them, Variant 3, contains sequence from the intron that precedes exon 4 (Figure 1, Figure 2). To simplify comparisons, the nomenclature used in this paper for labeling CNGA3 exons is in accordance with reference .
The novel exon observed in Variant 3 is 165 bases. It is preceded by a 257 base intron and followed by a 2022 base intron as identified in genomic sequence from chromosome 2q11 (OMIM 600053). This exon conveys an additional 55 in-frame codons to the cDNA (Figure 1). These extra amino acids are inserted into, and potentially may disrupt, the calmodulin binding site . The novel exon has conserved acceptor and donor ends for intron splicing  (Figure 3). However, it is not preceded by a polypyrimidine tract as is typically the case for constitutively expressed exons  and there is no branch point consensus sequence (YNYURAY) for the formation of a lariat intermediate loop between its 5' end and the preceding exon (exon 3). The absence of these signal sequences may be the cause of its not being a constitutive exon. An hCNGA3 cDNA previously identified from cone cells contains this exon . However, the Variant 3, from testes, differs from this form in that Variant 3 does not contain exon 5.
CNGA3 transcripts were detected in all tissues examined (Figure 4). The expected RT-PCR product sizes from the three forms of hCNGA3 are shown in Table 2. In order to confirm product identity, DNA from sample bands was purified from the gel, and sequence verified. The following bands were confirmed: the 168 and 333 bp bands from primer set A in testes; the 143 and 197 bp bands from primer set B in testes, cerebellum and pancreas; the 197 bp band from primer set B in retina. Based on their sequences these bands were, as expected, amplified from the alternatively spliced forms of hCNGA3. RT-PCR products from testes Quickclone cDNA (BD Biosciences) were not different from the products from testes cDNA prepared using Superscript II Reverse Transcriptase (Invitrogen).
RT-PCR using primer set A (Figure 4A) generated less intense bands from Variant 3 transcripts than from those of Variant 2 and/or Variant 1. The 333 bp band resulting from Variant 3 can be detected, albeit at lower intensity levels, in 21 of the 23 tissues examined with exceptions being dorsal root ganglion and liver. The more intensely amplified 168 bp product in each tissue is from either Variant 2 or Variant 1, as these two forms cannot be distinguished using this primer set. Relative expression levels of variants within each tissue can be reasonably inferred from band intensities. Samples from each PCR reaction were taken from earlier cycles, 25, 30, and 35, for comparison to the final band intensities. Although proportional increases in band intensities were seen with increasing cycle number, relative intensities among bands did not change (data not shown). This indicates a relative excess of those transcripts yielding brighter bands, although more sophisticated methods for transcript quantification are needed to directly address this aspect.
Primer set B was designed to distinguish the hCNGA3 Variant 1 from Variant 2. Bands representing Variant 2 are brightest in every sample except retina. In retina, the transcript containing exon 5 (Variant 1) yielded the brightest band (compare bands at 143 bp and 197 bp). Retina is the only sample in which robust amplification of Variant 1 transcript was observed. Within its genomic context, the exon 5, which defines Variant 1, has conserved donor and acceptor splice sites. In addition, it is preceded by an intronic canonical branch point sequence (CCTTGAT) for the facilitation of the lariat intermediate loop and a typical polypyrimidine tract. No aberrant or weak signal sequences are present that may contribute to its being ignored by the spliceosome as a constitutive exon. Other less well defined sequence signals may be present that play a role in silencing and/or enhancing the expression of exon 5 (see Discussion).
Attempts were made to identify the RT-PCR products with greater than expected size. DNA was purified from gel slices taken from these regions and cloned into pcDNA3.1/V5-His-TOPO (Invitrogen). One band (amplified from testes cDNA) was identified as a 280 bp product, generated by primer set B, representing premature CNGA3 mRNA; this product contained sequence from the intron preceding exon 5 that disrupted the reading frame resulting in a premature stop signal. None of the other attempts at cloning products with greater than expected size was successful due to low yields from the gel purification.
In the nervous system, tissue specific expression of alternatively spliced genes is not uncommon . Moreover, in a recent genome wide survey of alternatively spliced transcripts from human tissues, retina had one of the highest incidences with twice the number of tissue specific splice forms than the average tissue . Accordingly, the identification of alternatively spliced transcripts of a gene, which has hitherto only been described from the retina, may not be unexpected. Indeed, cDNA comparison of CNGA3 from various species and tissues reveals a variety of forms that differ in their exonic constitution . However, detailed studies on splice forms of human CNGA3 have not been undertaken thus far.
In the human cone CNGA3 (Variant 1), exons 5 and 7 are adjacent, with no intervening sequence homologous to the exon 6, which has been identified in other species. Variants 2 and 3, presented here, are missing both exons 5 and 6 so that exons 4 and 7 are adjacent. An analogous pattern of differential splicing has been shown in CNGA3 from chicken. The CNGA3 from chicken retina contains exons 1 through 9 ; however two variants of CNGA3 were also isolated from the chicken pineal organ missing exon 6 or both exons 5 and 6 . RT-PCR demonstrated that the shorter transcript (missing both exons 5 and 6) was the singular form of CNGA3 in chicken testes RNA . This is similar to our findings in human tissues in that this short form (Variant 2) is the predominant form in testes.
Functional expression of the shortened chicken CNGA3 demonstrated a two fold higher sensitivity to cNMPs compared to the variant from retina, which has exons 5 and 6 . The human CNGA3 exon 5, as it is expressed in cone cells, encodes 18 amino acids for which a function has yet to be inferred. There is some evidence that this region of the protein masks the potential for calmodulin (CaM) modulation. As demonstrated by Grunwald et al. , the human CNGA3 homomeric channels from cone are not capable of CaM modulation in spite of the fact that there is a CaM binding domain present in the protein encoded by exon 4. Through mutational analysis of the region surrounding exon 4 it was shown that portions of the protein encoded for by exon 5 may play a role in suppressing CaM modulation. Perhaps, the different expression of CNGA3, as presented here, is an indicator that the suppression of CaM modulation is essential for the proper function of CNGA3 in cone and not in other tissues, although functional studies are warranted to address this further.
Given its conserved signal sequences for splicing, the alternative expression of exon 5 is potentially due to enhancer and/or silencer sequences located in or around it. The interaction between exonic splicing enhancer (ESE) sequences and Serine/Argenine rich (SR) nuclear proteins has been shown to play a significant role in the recognition, by the spliceosome, of constitutive exons among myriad intronic possibilities . These ESE sequences are short and degenerate, thus they are difficult to identify without empirical evidence. Using ESEfinder 2.0 , eleven potential ESE sequences can be identified in exon 5, only 54 bases long. The ESEfinder 2.0 is designed to identify ESE sequences for four different SR proteins. Other SR proteins exist for which there may be binding sites within exon 5; binding sites of these are not well established. Exonic splicing silencer (ESS) sequences are, as yet, less defined than ESE sequences [30,31]. Motifs previously identified as ESS sequences were searched for in exon 5 and none were found.
Since exon 5 is seemingly favored as a constitutive exon in the retina, its splicing is enhanced and/or its silencing is disrupted in this tissue. Other retina specific alternatively spliced exons were compared with the CNGA3 exon 5 for the purpose of identifying commonalities that may be significant for their enhanced retinal expression. Exon 15a of retinitis pigmentosa GTPase regulator (RPGR) , insert A and B of the amphiphysin I , and exon 9a of metabotrobic glutamate receptor 7 (GRM7)  were all examined for similarities to the hCNGA3 exon 5 with respect to splicing signals. No common motifs were identified. It is likely that mechanisms responsible for the retina specific selection of these exons are complex enough to render simple primary sequence analysis insufficient for their proper identification, given the present understanding.
The purine rich sequence at the 3' end of exon 5 may play a role in its splicing regulation as a number of purine rich splicing regulators have been identified . CNGA3 exon 6, as it is expressed in non-human homologs, is purine rich and, as mentioned, has shown tissue specific alternative expression. However, hCNGA3 exon 4 also has a purine rich 3' end and is constitutively expressed regardless of tissue type.
Our study revealed a widespread distribution of CNGA3 RNA among human tissues. Previous studies designed to detect CNGA3 transcripts using Northern analysis, RT-PCR or in situ hybridization have been carried out with tissues from bovine, chicken, rat and rabbit. However, these studies have not extensively examined CNGA3 distribution in diverse tissues. CNGA3 transcripts have been detected in testes, retina, and pineal tissues from bovine and chicken in addition to bovine kidney, heart, and adrenal gland. Transcripts of CNGA3 were not detectable in bovine cerebellum, ileum, or tongue [5,14].
In the present study, we have isolated two unique transcripts of the human CNG channel alpha subunit, hCNGA3. Variant 2 and Variant 3, differ from the originally described form, Variant 1 from cone, in that they lack sequence from exon 5. Variant 3 differs from Variant 2 in that it contains an extra exon encoding 55 amino acids between exons 3 and 4. In results from the RT-PCR analysis of various tissues, bands representing Variant 2 were most intense in all samples, except retina. The unique predominance of Variant 1 in the retina implies a difference between the functional role of CNGA3 in cone cells from that in other tissues. Additionally, these findings point to an important role for the sequence encoded by exon 5 in cone cells. Its widespread distribution among human tissues indicates a broader role for CNGA3. Further studies could help elucidate physiological functions of hCNGA3 splice forms and may shed light on the basis for their tissue specific expression.
1. Kaupp UB, Seifert R. Cyclic nucleotide-gated ion channels. Physiol Rev 2002; 82:769-824.
2. Dryja TP, Finn JT, Peng YW, McGee TL, Berson EL, Yau KW. Mutations in the gene encoding the alpha subunit of the rod cGMP-gated channel in autosomal recessive retinitis pigmentosa. Proc Natl Acad Sci U S A 1995; 92:10177-81.
3. Bareil C, Hamel CP, Delague V, Arnaud B, Demaille J, Claustres M. Segregation of a mutation in CNGB1 encoding the beta-subunit of the rod cGMP-gated channel in a family with autosomal recessive retinitis pigmentosa. Hum Genet 2001; 108:328-34.
4. Paloma E, Martinez-Mir A, Garcia-Sandoval B, Ayuso C, Vilageliu L, Gonzalez-Duarte R, Balcells S. Novel homozygous mutation in the alpha subunit of the rod cGMP gated channel (CNGA1) in two Spanish sibs affected with autosomal recessive retinitis pigmentosa. J Med Genet 2002; 39:E66.
5. Distler M, Biel M, Flockerzi V, Hofmann F. Expression of cyclic nucleotide-gated cation channels in non-sensory tissues and cells. Neuropharmacology 1994; 33:1275-82.
6. Feng L, Subbaraya I, Yamamoto N, Baehr W, Kraus-Friedmann N. Expression of photoreceptor cyclic nucleotide-gated cation channel alpha subunit (CNGCalpha) in the liver and skeletal muscle. FEBS Lett 1996; 395:77-81.
7. Ding C, Potter ED, Qiu W, Coon SL, Levine MA, Guggino SE. Cloning and widespread distribution of the rat rod-type cyclic nucleotide-gated cation channel. Am J Physiol 1997; 272:C1335-44.
8. Sautter A, Zong X, Hofmann F, Biel M. An isoform of the rod photoreceptor cyclic nucleotide-gated channel beta subunit expressed in olfactory neurons. Proc Natl Acad Sci U S A 1998; 95:4696-701.
9. Biel M, Seeliger M, Pfeifer A, Kohler K, Gerstner A, Ludwig A, Jaissle G, Fauser S, Zrenner E, Hofmann F. Selective loss of cone function in mice lacking the cyclic nucleotide-gated channel CNG3. Proc Natl Acad Sci U S A 1999; 96:7553-7.
10. Sundin OH, Yang JM, Li Y, Zhu D, Hurd JN, Mitchell TN, Silva ED, Maumenee IH. Genetic basis of total colourblindness among the Pingelapese islanders. Nat Genet 2000; 25:289-93.
11. Kohl S, Marx T, Giddings I, Jagle H, Jacobson SG, Apfelstedt-Sylla E, Zrenner E, Sharpe LT, Wissinger B. Total colourblindness is caused by mutations in the gene encoding the alpha-subunit of the cone photoreceptor cGMP-gated cation channel. Nat Genet 1998; 19:257-9.
12. Johnson S, Michaelides M, Aligianis IA, Ainsworth JR, Mollon JD, Maher ER, Moore AT, Hunt DM. Achromatopsia caused by novel mutations in both CNGA3 and CNGB3. J Med Genet 2004; 41:e20.
13. Bonigk W, Altenhofen W, Muller F, Dose A, Illing M, Molday RS, Kaupp UB. Rod and cone photoreceptor cells express distinct genes for cGMP-gated channels. Neuron 1993; 10:865-77.
14. Biel M, Zong X, Distler M, Bosse E, Klugbauer N, Murakami M, Flockerzi V, Hofmann F. Another member of the cyclic nucleotide-gated channel family, expressed in testis, kidney, and heart. Proc Natl Acad Sci U S A 1994; 91:3505-9.
15. Bonigk W, Muller F, Middendorff R, Weyand I, Kaupp UB. Two alternatively spliced forms of the cGMP-gated channel alpha-subunit from cone photoreceptor are expressed in the chick pineal organ. J Neurosci 1996; 16:7458-68.
16. Timpe LC, Jin KL, Puelles L, Rubenstein JL. Cyclic nucleotide-gated cation channel expression in embryonic chick brain. Brain Res Mol Brain Res 1999; 66:175-8.
17. Meyer MR, Angele A, Kremmer E, Kaupp UB, Muller F. A cGMP-signaling pathway in a subset of olfactory sensory neurons. Proc Natl Acad Sci U S A 2000; 97:10595-600.
18. Weyand I, Godde M, Frings S, Weiner J, Muller F, Altenhofen W, Hatt H, Kaupp UB. Cloning and functional expression of a cyclic-nucleotide-gated channel from mammalian sperm. Nature 1994; 368:859-63.
19. Wiesner B, Weiner J, Middendorff R, Hagen V, Kaupp UB, Weyand I. Cyclic nucleotide-gated channels on the flagellum control Ca2+ entry into sperm. J Cell Biol 1998; 142:473-84.
20. Yu WP, Grunwald ME, Yau KW. Molecular cloning, functional expression and chromosomal localization of a human homolog of the cyclic nucleotide-gated ion channel of retinal cone photoreceptors. FEBS Lett 1996; 393:211-5.
21. Wissinger B, Muller F, Weyand I, Schuffenhauer S, Thanos S, Kaupp UB, Zrenner E. Cloning, chromosomal localization and functional expression of the gene encoding the alpha-subunit of the cGMP-gated channel in human cone photoreceptors. Eur J Neurosci 1997; 9:2512-21.
22. Grunwald ME, Zhong H, Lai J, Yau KW. Molecular determinants of the modulation of cyclic nucleotide-activated channels by calmodulin. Proc Natl Acad Sci U S A 1999; 96:13444-9.
23. Stephens RM, Schneider TD. Features of spliceosome evolution and function inferred from an analysis of the information at human splice sites. J Mol Biol 1992; 228:1124-36.
24. Cartegni L, Chew SL, Krainer AR. Listening to silence and understanding nonsense: exonic mutations that affect splicing. Nat Rev Genet 2002; 3:285-98.
25. Wissinger B, Gamer D, Jagle H, Giorda R, Marx T, Mayer S, Tippmann S, Broghammer M, Jurklies B, Rosenberg T, Jacobson SG, Sener EC, Tatlipinar S, Hoyng CB, Castellan C, Bitoun P, Andreasson S, Rudolph G, Kellner U, Lorenz B, Wolff G, Verellen-Dumoulin C, Schwartz M, Cremers FP, Apfelstedt-Sylla E, Zrenner E, Salati R, Sharpe LT, Kohl S. CNGA3 mutations in hereditary cone photoreceptor disorders. Am J Hum Genet 2001; 69:722-37.
26. Grabowski PJ, Black DL. Alternative RNA splicing in the nervous system. Prog Neurobiol 2001; 65:289-308.
27. Xu Q, Modrek B, Lee C. Genome-wide detection of tissue-specific alternative splicing in the human transcriptome. Nucleic Acids Res 2002; 30:3754-66.
28. Liu HX, Zhang M, Krainer AR. Identification of functional exonic splicing enhancer motifs recognized by individual SR proteins. Genes Dev 1998; 12:1998-2012.
29. Cartegni L, Wang J, Zhu Z, Zhang MQ, Krainer AR. ESEfinder: A web resource to identify exonic splicing enhancers. Nucleic Acids Res 2003; 31:3568-71.
30. Zhang XH, Chasin LA. Computational definition of sequence motifs governing constitutive exon splicing. Genes Dev 2004; 18:1241-50.
31. Sironi M, Menozzi G, Riva L, Cagliani R, Comi GP, Bresolin N, Giorda R, Pozzoli U. Silencer elements as possible inhibitors of pseudoexon splicing. Nucleic Acids Res 2004; 32:1783-91.
32. Kirschner R, Rosenberg T, Schultz-Heienbrok R, Lenzner S, Feil S, Roepman R, Cremers FP, Ropers HH, Berger W. RPGR transcription studies in mouse and human tissues reveal a retina-specific isoform that is disrupted in a patient with X-linked retinitis pigmentosa. Hum Mol Genet 1999; 8:1571-8.
33. Terada Y, Tsutsui K, Sano K, Hosoya O, Ohtsuki H, Tokunaga A, Tsutsui K. Novel splice variants of amphiphysin I are expressed in retina. FEBS Lett 2002; 519:185-90.
34. Schulz HL, Stohr H, Weber BH. Characterization of three novel isoforms of the metabotrobic glutamate receptor 7 (GRM7). Neurosci Lett 2002; 326:37-40.
35. Elrick LL, Humphrey MB, Cooper TA, Berget SM. A short sequence within two purine-rich enhancers determines 5' splice site specificity. Mol Cell Biol 1998; 18:343-52.