- Email: [email protected]
A-to-I pre-mRNA editing of the serotonin 2C receptor: Comparisons among inbred mouse strains
Gene 382 (2006) 39 – 46 www.elsevier.com/locate/gene
A-to-I pre-mRNA editing of the serotonin 2C receptor: Comparisons among inbred mouse strains Yunzhi Du a,b , Muriel T. Davisson e , Karen Kafadar d , Katheleen Gardiner a,b,c,⁎ a
Eleanor Roosevelt Institute at the University of Denver, USA Program in Human Medical Genetics, University of Colorado at Denver and the Health Sciences Center, Denver, Colorado, USA Department of Biochemistry and Molecular Biology, University of Colorado at Denver and the Health Sciences Center, Denver, Colorado, USA d Department of Mathematics, University of Colorado at Denver and the Health Sciences Center, Denver, Colorado, USA e The Jackson Laboratory, Bar Harbor, Maine, USA b
c
Received 20 April 2006; received in revised form 30 May 2006; accepted 7 June 2006 Available online 27 June 2006
Abstract The serotonin receptor 5HT2CR pre-mRNA is subject to adenosine deamination (RNA editing) at five residues located within a 15 nucleotide stretch of the coding region. Such changes of adenosine to inosine (A-to-I) can produce 32 mRNA variants, encoding 24 different protein isoforms, some of which vary in biochemical and pharmacological properties. Because serotonin mediates diverse neurological processes relevant to behavior and because inbred mouse strains vary in their responses to tests of learning and behavior, we have examined the A-to-I editing patterns of the 5HT2CR mRNA in whole brains from eight mouse strains. By sequencing approximately 100 clones from individual mice, we generated detailed information on levels of editing at each site and patterns of editing that identify a total of 28 mRNA and 20 protein isoforms. Significant differences between individuals from different strains were found in total editing frequency, in the proportion of transcripts with 1 and 4 edited sites, in editing frequency at the A, B, E and D sites, in amino acid frequencies at positions 157 and 161, and in subsets of major protein isoforms. Primer extension assays were used to show that individuals within strains (six C3H.B−+rd1 and four 129SvImrJ) displayed no significant differences in any feature. These findings suggest that genetic background contributes to subtle variation in 5HT2CR mRNA editing patterns which may have consequences for pharmacological treatments and behavioral testing. © 2006 Elsevier B.V. All rights reserved. Keywords: 5HT2C receptor; Inosine; Protein isoforms; Cloning; Primer extension; Behavior
1. Introduction The serotonin receptor 2C (5HT2CR) is widely distributed within the central nervous system (Pompeiano et al., 1994; Liu et al., 1999). It mediates diverse neurological processes that affect feeding behavior, sleep, sexual behavior, anxiety and depression (Tecott et al., 1995; Brennan et al., 1997, reviewed by Graeff et al., 1996). The 5HT2CR pre-mRNA is the only G protein coupled receptor transcript known to be a target of site-specific adenosine deamination (A-to-I pre-mRNA editing). Editing can produce Abbreviations: 5HT2CR, serotonin receptor 2C; I, inosine; A, adenosine; G, guanosine. ⁎ Corresponding author. Eleanor Roosevelt Institute at the University of Denver, 1899 Gaylord Street, Denver, Colorado, 80206, USA. Tel.: +1 303 336 5652; fax: +1 303 333 8423. E-mail address: [email protected] (K. Gardiner). 0378-1119/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2006.06.007
inosine at five sites (A, B, E, C, D) within a 15 nucleotide segment located within the putative second intracellular loop (Burns et al., 1997). These editing sites are conserved between human and rodents, although predominant editing patterns and levels differ (Niswender et al., 1999). Because inosine is interpreted by the translation machinery as guanosine, editing can change the genomically encoded amino acid sequence 157-IRNPI-161 at I, N and I, as shown in Fig. 1. Considering editing independently at each site, 32 mRNA variants encoding 24 protein isoforms are theoretically possible (Burns et al., 1997). Editing of 5HT2CR mRNA affects biochemical and pharmacological properties of the protein. Of the isoforms tested, the unedited INI and the fully edited VGV isoforms display the extremes of several properties, with INI the most active isoform in the absence of agonist, and VGV the least, and INI displaying the highest agonist affinity and potency, and VGV the lowest (Niswender
40
Y. Du et al. / Gene 382 (2006) 39–46
TJL's IACUC. Eight-week old male mice from inbred strains: C57BL/6J, DBA/2J (DBA), C3H.B− +rd1 (C3H ), 129S1/SvImJ (129), FVB/NJ ( FVB), BUB/BnJ (BUB) and CAST/Ei (CAST ), and the F1 hybrid B6C3.b F1 (B6C) (obtained by crossing C57BL/6J and C3H.B− + rd1 ) were sacrificed by cervical dislocation, all between 9 and 11 AM. One mouse was used from each strain, except for two C57BL/6J, seven C3H and five 129. Brains were dissected out and immediately frozen in liquid nitrogen. Fig. 1. Edited region of the 5HT2C receptor. Nucleotides 1159–1173 of the mouse 5HT2CR mRNA (NM_008312) identical to the genomic sequence are shown. Adenosine residues that can be edited are labeled as sites A–E (site E is also known as site C′). The genomically encoded amino acid sequence from positions 157–161, is IRNPI, abbreviated as INI in the completely unedited form. Amino acid changes corresponding to each editing event are indicated.
et al., 1999; Herrick-Davis et al., 1999). Other, partially edited, isoforms typically display intermediate characteristics. The predominant isoform in rodent brain, VNV, shows significantly decreased receptor basal activity compared with INI (Herrick-Davis et al., 1999), and VSV, MSV, IDV, IGV and VGI result in decreased inositol phosphate release in response to serotonin (Wang et al., 2000). Lastly, basal activity of INI is significantly reduced by exposure to the antipsychotic drugs loxapine and clozapine, while VGV is unaffected (Niswender et al., 2001), and, while INI and VGV bind lysergic acid diethylamide (LSD) with similar affinity, only VGV exhibited a markedly reduced response, as measured by phosphoinositide hydrolysis (Niswender et al., 2001). These data suggest that alteration of editing patterns or levels, whether by genetic, environmental or stochastic processes, may have neurological consequences affecting behavior and responses to the environment and pharmacological agents. Indeed, several recent reports have described statistically significant differences in editing patterns between controls and subsets of patients with depression (Gurevich et al., 2002a; Niswender et al., 2001), and between controls and mice treated with a 5HT2C agonist or antagonist (Gurevich et al., 2002a,b), including some mouse strain differences (Englander et al., 2005; Hackler et al., 2006). To further explore the genetic basis of 5HT2CR editing patterns, with a long term goal of uncovering additional correlations between 5HT2CR editing levels and behavior, we have examined editing patterns in whole brains from eight strains of mice. Individuals within an inbred strain are genetically identical, and members of different strains are known to differ in their responses to specific behavioral tests (reviewed in Crawley et al., 1997), including those designed to assess anxiety (Bouwknecht and Paylor, 2002). Comparison of editing pattern differences within and among strains suggests that there are biologically subtle, but statistically significant, genetic effects that may be relevant to 5HT2CR function. 2. Materials and methods 2.1. Mouse brain tissue All mice were raised and sacrificed at The Jackson Laboratory (TJL) under TJL animal protocol number 99092 approved by
2.2. Cloning and analysis of 5HT2CR sequences RNA was extracted from individual brains from each strain, using the standard guanidine-phenol method and RNA integrity was verified by gel electrophoresis. One microgram of total RNA was used in first-strand cDNA synthesis with MMLV reverse transcriptase and random hexamer priming (Invitrogen). The segment spanning the edited region was amplified by the polymerase chain reaction ( PCR) with primers 5′ TGT GCT ATT TTC AAC TGC GTC CAT CAT G 3′ and 5′ CGG CGT AGG ACG TAG ATC GTTAAG 3′ located in exon 5 and exon 6, respectively, of the mouse 5HT2CR mRNA sequence ( NM_008312). Fresh PCR product (327 bp) was cloned into PCRII 2.1 ( Invitrogen) and transformed into E. coli. For each brain, approximately 100 individual clones were picked from a single transformation. Inserts were sequenced on both strands using Big Dye Terminator 3.0 cycle sequencing kit (ABI) and M13 reverse and forward primers 5′ CAGGAAACAGCTATGAC3′ and 5′GTAAAACGACGGC CAG3′. Sequence traces were examined for the presence of an A or a G residue at each of the five editing sites, and the complete sequence of the edited region of each clone was scored. Sequences with high background or ambiguous base calling were discarded. Between 80 and 106 sequences were determined for each brain. 2.3. Primer extension assays Total RNA isolation, reverse transcription and PCR were performed as described above on six additional brains from C3H and four additional brains from 129 (these did not include samples used in cloning experiments). PCR products were purified through ChromaSpin TE100 columns (Clontech). Forty ng of PCR product was used as template for primer extension assays. Primers and assay conditions were modified from Chen et al. (2000) and Niswender et al. (2001). Forward strand primer extA18mer, 5′CGCTGGACCGGTATGTAG3′, was used to measure A site editing frequency in the presence of 4.3 mM ddTTP and ddGTP, 0.7 mM dATP and dCTP. Forward strand primer extA20mer, 5′CGCTGGACCGGTATGTAGCA3′, was used to measure B site editing frequency in the presence of 6 mM ddATP and 1 mM each dTTP, dGTP and dCTP or 6 mM ddGTP, and 1 mM dATP, dTTP and dCTP. Reverse strand primer extD, 5′ GAATTGAACCGGCTATGCTC3′, was used to measure D, C and E site editing frequencies in the presence of 6 mM ddTTP, 1 mM dATP, dGTP and dCTP or 6 mM ddCTP, 1 mM dATP, dTTP and dGTP. Annealing was carried out at 25 °C for 30 min, and extension, at 37 °C, for 10 min in the presence of 7 mM
Y. Du et al. / Gene 382 (2006) 39–46 Table 1 Site-specific editing frequencies
41
and ethanol precipitation. The region putatively involved in RNA duplex formation, containing the edited sites in exon 5 and the 5′ end of intron 5, was amplified by PCR using primers 5′ GTATGGTCAGTGCAATTTGGATAAC3′ and 5′GTAT CAGTGTTGCCAAAATCCACTG3′, located in introns 4 and 5, respectively, of the mouse genomic sequence ( NC_000086). The PCR product (518 bp) was sequenced directly using the same primers, as described above. 3. Results 3.1. Cloning and sequencing
Percentage of clones with a G residue at each of the sites A–E was calculated from the number of clones sequenced. Significant differences ( p < 0.05) are indicated in bold (high) and bold shaded (low); * trend towards significance (0.05 < p < 0.1). P values are indicated in the text. For comparison with primer extension data in Table 7, 95% confidence limits for A, B C and D sites are between 6% and 11%.
MnCl2, 10 mM DTT, 1× sequencing buffer, 1.5 units Sequenase version 2.0 (USB). Products were resolved on standard denaturing 15% polyacrylamide gels, exposed to a phosphor screen and scanned on a PhosphoImager (Molecular Dynamics, Strom 840). Signal intensities of bands and relative ratios were quantified using ImageQuant (Molecular Dynamics).
Between 81 and 106 clones per individual, for a total 740 clones, were sequenced from one individual from each of seven inbred strains and one F1 hybrid. An additional 29 clones were sequenced from a second individual from C57BL/6J. Each sequence was scored for the presence of a G residue, indicating editing, at each of the five sites, A–E, shown in Fig. 1. Sequences were used to calculate the following editing patterns in individual mice: the proportion of editing at individual sites, the proportion of editing at all five sites, the proportion of transcripts edited at different numbers of sites, the proportion of each mRNA isoform, the occurrence of specific amino acids at individual sites, and the occurrence of each protein isoform. Data for all mice are summarized in Tables 1–6. Pairwise comparisons were carried out between individual mice.
2.4. Statistical analyses 3.2. Variation in editing frequency at individual sites All statistical analysis used the GraphPad Prism4 software. Differences in editing frequency at five editing sites and in amino acid frequencies and protein isoforms obtained from sequence analysis of clones from individual mice were analyzed using Fisher's Exact Test to compare all possible pairs. The Modified Wold method was used to determine 95% confidence limits in individual site editing levels. Results of quantification of primer extension assays were analyzed with Two-Tailed Unpaired Students t-test. 2.5. Analysis of 5HT2CR genomic sequence Genomic DNA was isolated from the livers of each mouse analyzed in the cloning experiment using standard proteinase K digestion plus phenol chloroform extraction, RNase A digestion
Table 2 Percentage of transcripts with 0–5 inosines
Bold, shading and * as in Table 1.
As shown in Table 1, editing levels at the A site in C3H, BUB and B6C, at 89%, 87% and 81%, respectively, were significantly higher than those in DBA and 129, at 67% and 65% ( p < 0.001, p < 0.01 and p < 0.05). C3H and BUB were also significantly higher than FVB ( p < 0.01 and 0.05). Editing at the B site was significantly higher in BUB, C3H and B6C (73%–82%) relative to 129 (52%) ( p < 0.001, p < 0.0001 and p < 0.05, respectively). C3H was also higher than DBA ( p < 0.01). At the E site, C3H and BUB showed significantly higher levels (10% and 9%) than FVB (1%) ( p < 0.05) and a trend toward higher levels than CAST at 2% ( p = 0.06). No significant differences were found for the editing frequency at the C site. At the D site, B6C editing, at 79%, was significantly higher than CAST at 66% ( p < 0.05); C3H (79%)
Table 3 Frequencies of major mRNAs and unedited and completely edited variants
Bold, shading as in Table 1.
42
Y. Du et al. / Gene 382 (2006) 39–46
Table 4 Editing at A and B sites is not independent
*, + sign, denotes editing to an inosine residue; − sign, denotes an unedited, genomic A Residue. Subscripts refer to the editing status of the context site, i.e. A+B+ is the ratio of A+B+ to (A+B+ + A−B+), where edited B site is the context site. Bold, indicates all values in the column are significantly greater than values in the adjacent column in bold shaded.
showed a trend toward a significant difference from CAST ( p < 0.07). The last column of Table 1 shows the total amount of editing at all five sites. C3H, at 57%, had significantly higher total editing than DBA, 129, FVB, CAST and C57BL/6J at 44%– 49% ( p < 0.05 to p < 0.0001). BUB and B6C, at 52%, had higher levels than 129 (44%, p < 0.05). Cloning and sequencing were carried out for a smaller number of clones (29) from a second brain of C57BL/6J. Editing levels were not significantly different between the two mice (Table 1). 3.3. Variation in editing levels among transcripts The presence of inosine within a transcript may have consequences for translational efficiency, especially when there are multiple inosines within a short stretch. Table 2 shows the proportion of transcripts edited at any combination of 1 to 4 sites. C3H has the highest proportion of transcripts, 26%, edited at four sites, and the lowest proportion of transcripts, 7%, edited at a single site. 129, FVB and CAST show the opposite proportions, with only 10%–12% edited at four sites but 20%–22% edited at a single site. These differences are significant ( p < 0.05). Other significant differences in the level of transcripts with one site edited include C3H < DBA, BUB < 129, and B6C < 129 and CAST ( p < 0.05). C3H and 129 also show trends towards significant differences in the proportion of transcripts edited at three sites and two sites, with 50% and 13%, respectively in C3H, compared with 37% and 23%, in 129 ( p = 0.09 and p = 0.08, respectively). 3.4. Variation in mRNA editing patterns Differences in mRNA editing patterns may exist that are not apparent from editing frequencies at individual sites. In this analysis, a clone sequence labeled ABD indicates sites A, B and D were edited to inosines, while sites C and E remained the genomic A residues. Such combinations of editing theoretically can produce 32 different RNA sequences. The frequencies of the six most commonly observed variants are shown in Table 3. The frequencies of the fully edited and completely unedited forms are included for comparison. In all eight strains, the predominant form was ABD, ranging from 33% to 46% and
representing an average of 37% of mRNAs in all samples. The other common mRNA variants were ABCD with an average frequency of 12%, AB with 8%, D only with 10%, and AD, A only and the completely unedited form, each averaging ∼ 6%. The fully edited form was very rare, seen only in two mice and only at 1%. Four mRNA variants were not observed in the entire 740 clone set: BE, BEC, BED and BECD. The remaining 20 variants were observed in fewer than 5% of clones in each mouse. In comparing individual mice, the D site variant occurred less frequently in C3H (2%) than in 129, FVB, DBA and CAST (12%–17%, p < 0.01–0.05), and in BUB less often (4%) than in 129 ( p < 0.01). 3.5. Independence of editing at adjacent sites Enzymes that carry out editing, Adenosine Deaminases that Act on RNA, ADAR1 and ADAR2, display preferences in nearest neighbors surrounding editing sites (Lehmann and Bass, 2000; Dawson et al., 2004). Studies with mouse models with null mutations in one or both enzymes suggest that in 5HT2CR the A and B sites are predominantly edited by ADAR1 and the E, C and D sites, by ADAR2 (Higuchi et al., 2000; Wang et al., 2004; Hartner et al., 2004). Because the five 5HT2CR sites are so closely spaced, we asked if editing at one site influenced the observed frequency of editing at adjacent sites. For this, we calculated the percent editing at each site in the context of +/− editing at the immediately adjacent site. For example, A+B+ is
Table 5 Frequencies of cDNAs encoding major protein isoforms and V at 157
Bold, shading and * as in Table 1.
Y. Du et al. / Gene 382 (2006) 39–46
43
Table 6 Frequencies of cDNAs encoding minor protein isoforms
C57BL/6J DBA C3H 129 FVB BUB CAST B6C
ISI
ISV
IDI
IDV
IGI
IGV
VDI
VDV
VGI
VGV
MNI
MNV
MSI
MSV
Total
0% 2% 0% 1% 1% 0% 1% 0%
3% 3% 1% 5% 2% 0% 2% 2%
0% 0% 1% 0% 0% 1% 0% 0%
2% 0% 0% 1% 0% 2% 0% 0%
0% 0% 0% 1% 0% 0% 0% 0%
1% 0% 0% 0% 0% 0% 0% 1%
1% 0% 1% 0% 0% 2% 1% 0%
0% 2% 4% 2% 1% 3% 1% 6%
1% 1% 1% 1% 0% 0% 0% 1%
1% 0% 2% 0% 0% 0% 0% 1%
0% 1% 0% 0% 1% 0% 0% 0%
0% 3% 1% 1% 4% 1% 2% 1%
0% 1% 0% 0% 0% 0% 0% 0%
0% 0% 1% 0% 0% 0% 0% 0%
9% 13% 12% 12% 9% 9% 7% 12%
the percent A sites edited in the context of B site edited, i.e. the ratio of A+B+ to (A+B+ + A−B+), where + denotes an I residue and − denotes a genomic A residue. Similarly, A+B− is the ratio of A+B− to (A+B− + A−B−). Results in Table 4 assess the independence of editing at sites A and B, E and C, and C and D. Editing at the A and B sites (which are separated by a single U residue) is not independent; if either A or B is edited, it is highly likely that the second site is also edited (p < 0.0001, paired Student t-test); this may in part be a reflection of both sites being edited by ADAR1. In contrast, editing at the E site (immediately 5′ to the C site) is very low regardless of C site status, and editing events at the C and D sites (separated by four nucleotides) are independent of each other. 3.6. Prediction of consequences for protein isoforms Editing at any of the five sites alone causes a codon change as indicated in Fig. 1. It is the protein coding consequences that will provide information on the functional implications of 5HT2CR mRNA editing. While 28 mRNA isoforms were observed, these coded for 20 of the 24 possible protein isoforms. Table 5 shows the frequencies of cDNAs encoding the six most common protein isoforms found and Table 6, the frequencies of the 14 most rarely observed cDNAs. These data allow direct extrapolation to the 5HT2CR isoform population, if it is assumed that all cDNA variants are transported and translated with uniform efficiency. Editing at the A site, +/− editing at the B site, changes the genomically-encoded isoleucine (I) to valine (V), and therefore the significant differences in A site editing noted above would be reflected in the frequency of protein isoforms containing V at position 157. At position 159, C3H and CAST had the highest percentage of clones (31%) that had been edited to encode S, D or G, however, these are not significantly different from the lowest (21%) seen in 129. Based on differences in editing at the D site, a significant difference in the presence of V at position 161 is predicted in comparing B6C with CAST, and a trend towards significance in comparing C3H and CAST (Table 1). At the individual isoform level, the variation in the frequency of D only isoform, noted above, would result in the same variations among strains of the frequency of the INV isoform (Table 5). Significantly higher levels of VSI (ABC and AC) are predicted in CAST (12%) relative to C3H (2%) ( p < 0.05), 129 (1%) ( p < 0.01) and B6C (3%) ( p < 0.05). Fourteen cDNA variants that encode minor protein isoforms are observed on average at < 1% in all strains. These are the
forms that contain methionine (M), and a subset of those containing glycine (G) and aspartic acid (D), and result from editing at B but not A, editing at E plus C, and editing at E but not C, respectively (Fig. 1). Table 6 lists the frequencies of these isoforms in the eight brains; typically fewer than half are represented in any sample, in total representing less than 13% of all isoforms. The four forms encoding Met plus Gly or Asp (MGV/I and MDV/I) were never observed. Lastly, although functional information is incomplete, subsets of isoforms can be grouped based on some biochemical properties. Herrick-Davis et al. (1999) measured basal activities of seven isoforms and ranked them as follows: if INI is assigned the maximal basal activity of 100%, ISV, INV and VSI display ∼ 75%, VNV and VSV, ∼ 50% and VGV, ∼ 22% of the maximal. The eight strains do not differ in levels of INI or VGV, but when isoforms of intermediate activities are grouped, several significant differences are seen (Table 5, columns 8,9). In particular, C3H has the highest levels (67%) of the less active VNV/VSV group and the lowest levels (5%) of the more active INV/ISV/VSI group. In contrast, DBA, 129 and CAST all have both lower levels (48%–50%) of the less active and higher levels (19%–26%) of the more active isoforms. Based on these data, C3H would be predicted to have overall less basal 5HT2CR activity compared with DBA, 129 and CAST (p < 0.05). In considering only the more active group, 129 has higher levels than BUB; CAST has higher levels than BUB and B6C; and FVB and C57BL/6J have higher levels than C3H. 3.7. Editing frequencies in C3H and 129 individuals Because all individuals from an inbred mouse strain are genetically identical to each other, if 5HT2CR editing patterns are solely genetically determined, no significant differences are expected among individuals within a strain. To examine individual variation, primer extension assays were used to measure editing frequencies in six additional individuals from C3H and four additional individuals from 129. These strains were chosen because they displayed the greatest differences in results from cloning and sequencing experiments. Representative gels are shown in Fig. 2 and results are summarized in Table 7. No significant differences in editing levels are seen within either strain, i.e. either among the six C3H individuals or among the four 129 individuals. In addition, at the A and B sites in 129 and at the C and D sites in both strains, levels of editing fall
44
Y. Du et al. / Gene 382 (2006) 39–46
Fig. 2. Representative primer extension gels. Primer extension assays were carried out on cDNA amplified from total brain RNA prepared from C3H and 129 individuals (#1 and #2), as described in the Materials and methods. (a) primer extA-18mer with ddGTP plus ddTTP, (b) primer extA-20mer with ddATP and ddGTP reactions (A and G, respectively), (c) primer extD with ddTTP and ddCTP reactions (T and C, respectively). Bands are identified by editing status (+, edited; −, unedited) and site (A, B, D, C or E); sizes of bands are indicated in nucleotides. P, primer.
within the 95% confidence limits of the cloning results shown in Table 1 (levels of editing at the A and B sites have been reported to be artefactually low by this method (Burns et al., 1997) and this may account for the small discrepancy in C3H from the values obtained by cloning). Importantly, editing at the A and B sites in C3H at 78% and 65%, respectively, is significantly higher than that in 129 at 70% and 54%, consistent with the cloning data. 3.8. Sequence of exon 5/intron 5 A-to-I pre-mRNA editing requires a double stranded RNA structure as a substrate (Bass, 2002; Maas et al., 2003; Seeburg and Hartner, 2003). In 5HT2CR pre-mRNA, this is an imperfect helix formed between the segment of exon 5 containing the edited region and the 5′ end of the downstream intron 5 (Burns et al., 1997). Variations in editing levels could be caused by polymorphisms that alter the stability of the double stranded structure so that enzyme binding is perturbed. To address this
Table 7 Editing frequency of individuals from C3H and 129 strains
Measured by primer extension as in Fig. 2.
possibility, the genomic sequence of exon five and the 5′ end of intron 5 encompassing the double stranded region was determined. All mice were homozygous in this region, and the nucleotide sequences were identical to previously published sequences (acc# NC_000086; data not shown). 4. Discussion In designing these experiments, we decided to examine individual mice and to use two approaches to determining 5HT2CR editing patterns. First, we used the more laborious, but more informative, approach of cloning cDNAs spanning the 5HT2C receptor editing site region and sequencing a large number of clones. This approach allowed us to determine frequencies of complete protein isoforms, to observe more rare protein isoforms, and to achieve statistical significance in several observed differences. Sequencing of a smaller number of clones from a second individual of one strain showed these two individuals did not differ significantly in editing levels and patterns. We then used the more rapid primer extension method to measure editing patterns of additional individuals from two different strains. Primer extension yields less detailed information on editing patterns and does not provide complete isoform information. However, statistical analysis of these data also indicated that individuals within a single strain did not differ significantly in their editing levels. This suggests the predominance of genetic regulation of editing over potential environmental and stochastic events. Genomic sequencing of the region believed to be involved in editing site selection showed no differences among strains, and therefore other mechanisms, such as activity levels of ADAR1 or ADAR2 that carry out editing, may be responsible for the inter-strain differences in editing levels. Based on these data, however, we conclude that it is reasonable to use the data from the cloning experiments in further strain comparisons. Sequencing 740 clones from the eight individual mouse brains identified a total of 28 of a possible 32 mRNA variants, encoding 20 of a possible 24 protein isoforms. This describes the most complete repertoire of isoforms so far observed in
Y. Du et al. / Gene 382 (2006) 39–46
mammalian brain. Only six isoforms were common and seen in all individuals (Table 5). The remaining 14 isoforms were rare (Table 6); only 5–8 of these were seen in any one brain and several were seen only in 1%–2% of sequences in only one or two brains. These general results are consistent with previous analyses in rodent brain (Burns et al., 1997; Niswender et al., 1999). Statistically significant differences in the frequency of editing at the A, B, E and D sites, in the total amount of editing, in the frequency of protein isoforms with V at position 157 and 161, and in the INV and VSI isoforms were observed in subsets of pairwise comparisons between strains. While whole brains were used here, it is known that editing levels vary, sometimes dramatically, among brain regions in humans and rodents (Niswender et al., 2001; Wang et al., 2000; Burns et al., 1997). Therefore, it is possible that some isoforms may be much more common and/or that variations between strains may be different when separate brain regions are examined, for example as has been shown in amygdala from C57BL/6J compared with DBA (Hackler et al., 2006). Biochemical and pharmacological consequences of editing have been assessed for some of the 24 isoforms. Most data have been generated for the completely unedited isoform, INI, the completely edited form, VGV, and occasionally for some partially edited forms (Berg et al., 2001; Price et al., 2001; Price and Sanders-Bush, 2000; Fitzgerald et al., 1999; McGrew et al., 2004; Marion et al., 2004). INI and VGV are too rare in whole brain to evaluate variation in their levels. However, for INV, ISV and VSI, Herrick-Davis et al. (1999) showed that basal activities were reduced by approximately 25% compared with INI. This is a less dramatic reduction than the approximately 50% seen with VNV and VSV. The proportions of isoforms with these relatively high and low basal activities are significantly different in C3H, and predict that C3H will have overall low basal activity relative to 129, DBA and CAST (Table 5). It would be useful to extend comprehensive isoform-specific comparisons to other biochemical and pharmacological properties, including coupling to different G proteins, activation of phospholipase A2 (PLA2) versus phospholipase C (PLC) for production of arachidonic acid and phosphoinositol, and responses to therapeutic and recreational drugs. If the INV/ISV/VSI isoform group typically differs from the VSV/VNV group in the strength of functional properties, these strains may differ in additional specific features of serotonin signaling. It is also of interest to consider protein levels of edited isoforms. Indirect evidence suggests that inosine in mRNA may destabilize translation through altered secondary structures or altered base pairing efficiency of tRNAs with inosine-containing codons. Most studies examined inosine in the 5′ UTR of mRNA or in the wobble position of tRNA (Kozak, 1980; Seal et al., 1989; Lim, 1995). Consequences may be significant in coding regions for first and second codon positions which occur in the A, E, C and D sites of 5HT2CR. Total editing levels of 5HT2CR vary among strains, with C3H displaying the highest levels, at 57%, compared to DBA, 129 and CAST (Table 1). In addition, however, C3H has the lowest proportion of transcripts with one inosine and the highest with four inosines (Table 2). If four inosines are relatively inefficiently translated, C3H may
45
have an isoform profile at the protein level more similar to 129, DBA and CAST, and/or may have lower total 5HT2CR protein levels. Editing is not a static feature. Yang et al. (2004) showed that treatment of human glioblastoma cell lines with Interferon-α resulted in dramatic increases in editing at the A, B and C sites (from <40% to >65%), and in the proportion of VSI isoform (from <10% to >50%). Interferon-α is used in treatment of some malignancies and has been associated with onset of depression (Schaefer et al., 2002). Gurevich et al. (2002a,b) showed that treatment of mice with fluoxetine (Prozac), a competitive antagonist of 5HT2CR, decreased editing at the E site from 9% to 1% and at the C site from 30% to 20%. Depletion of 5HT by treatment with an irreversible inhibitor of tryptophan hydroxylase produced similar results. Consistent with these data, treatment with the 5HT2CR agonist DOI increased editing at the E site. The authors concluded that editing levels are being modulated to regulate sensitivity to 5HT as environmental factors change. When initial levels of editing differ among individuals, however, it is necessary to determine if drug treatments, or environmental factors, maintain, abolish or otherwise modulate such differences. This is an important consideration for human treatments, where it has been shown that levels of editing, both site and isoform specific, vary considerably among normal controls (Gurevich et al., 2002a; Dracheva et al., 2003; Iwamoto and Kato, 2003). Comparisons among inbred strains, such as 129S1/SvImJ, FVB/NJ and C3H.B− +rd1, that normally differ in editing features affected by interferon and fluoxetine, will be useful for these experiments. However, it will also be necessary to consider that the predominant isoforms seen in human brain (e.g. VSV, INI and VNV at 38%, 10% and 8%) (Niswender et al., 2001; see also Fitzgerald et al., 1999) differ from those in rodent brain shown in Table 4 (see also Burns et al., 1997; Wang et al., 2000). Differences in editing patterns may also be relevant to responses to stressful environments. For example, strains 129 and C3H differ in responses to some tests associated with anxiety and also differ in levels of some significant isoforms. In context fear conditioning and the acoustic startle response, 129 performance is described as good and poor, respectively, and C3H as poor and moderate, while both are described as performing poorly in the open field locomotion test (Crawley et al., 1997). Bouwknecht and Paylor (2002) compared nine strains in stress-induced hyperthermia and exploration in the light–dark paradigm, tests which assay physiological and behavioral responses to stress, respectively. Interestingly, behavioral responses did not correlate with physiological responses, implicating multiple mechanisms. 5HT2CR editing patterns certainly would not be the sole factor governing stress or drug responses, but if they change with behavioral tests, particularly in specific cell types and/or brain regions, it will be important to determine if correlations exist between naïve editing patterns, and changes induced by stress and performance in aversive learning and memory tests. Acknowledgements This work was supported by grants from the Cullpepper Foundation and the National Institutes of Health to KG. The
46
Y. Du et al. / Gene 382 (2006) 39–46
authors thank Cecilia Schmidt and Dobromir Slavov for technical assistance. References Bass, B.L., 2002. RNA editing by adenosine deaminases that act on RNA. Ann. Rev. Biochem. 71, 817–846. Berg, K.A., Cropper, J.D., Niswender, C.M., Sanders-Bush, E., Emeson, R.B., Clarke, W.P., 2001. RNA-editing of the 5-HT(2C) receptor alters agonist– receptor–effector coupling specificity. Br. J. Pharmacol. 134, 386–392. Bouwknecht, J.A., Paylor, R., 2002. Behavioral and physiological mouse assays for anxiety: a survey in nine mouse strains. Behav. Brain Res. 136, 489–501. Brennan, T.J., Seeley, W.W., Kilgard, M., Schreiner, C.E., Tecott, L.H., 1997. Sound-induced seizures in serotonin 5HT2CR receptor mutant mice. Nat. Genet. 16, 387–390. Burns, C.M., et al., 1997. Regulation of serotonin-2C receptor G-protein coupling by RNA editing. Nature 387, 303–308. Chen, C.X., Cho, D.S., Wang, Q., Lai, F., Carter, K.C., Nishikura, K., 2000. A third member of the RNA-specific adenosine deaminase gene family, ADAR3, contains both single- and double-stranded RNA binding domains. RNA 6, 755–767. Crawley, J.N., et al., 1997. Behavioral phenotypes of inbred mouse strains: implications and recommendations for molecular studies. Psychopharmacology 132, 107–124. Dawson, T.R., Sansam, C.L., Emeson, R.B., 2004. Structure and sequence determinants required for the RNA editing of ADAR2 substrates. J. Biol. Chem. 279, 4941–4951. Dracheva, S., Elhakem, S.L., Marcus, S.M., Siever, L.J., McGurk, S.R., Haroutunian, V., 2003. RNA editing and alternative splicing of human serotonin 2c receptor in schizophrenia. J. Neurochem. 87, 1402–1412. Englander, M.T., Dulawa, S.C., Bhansali, P., Schmauss, C., 2005. How stress and fluoxetine modulate serotonin 2C receptor pre-mRNA editing. J. Neurosci. 25, 648–651. Fitzgerald, L.W., et al., 1999. Messenger RNA editing of the human serotonin 5HT2C receptor. Neuropsychopharmacology 21, 82S–90S. Graeff, F.G., Guimaraes, F.S., Andrade, T.G.C.S.D., Deakin, J.F.W., 1996. Role of 5HT in stress, anxiety, and depression. Pharmacol. Biochem. Behav. 54, 129–141. Gurevich, I., Tamir, H., Arango, V., Dwork, A.J., Mann, J.J., Schmauss, C., 2002a. Altered editing of serotonin 2c receptor pre-mRNA in the prefrontal cortex of depressed suicide victims. Neuron 34, 349–366. Gurevich, I., Englander, M.T., Adlersberg, M., Siegal, N.B., Schmauss, C., 2002b. Modulation of serotonin 2C receptor editing by sustained changes in serotonergic neurotransmission. J. Neurosci. 22, 10529–10532. Hackler, E.A., Airey, D.C., Shannon, C.C., Sodhi, M.S., Sanders-Bush, E., 2006. 5-HT(2C) receptor RNA editing in the amygdala of C57BL/6J, DBA/2J, and BALB/cJ mice. Neurosci. Res. 55, (Electronic publication ahead of print). Hartner, J.C., Schmittwolf, C., Kispert, A., Muller, A.M., Higuchi, M., Seeburg, P.H., 2004. Liver disintegration in the mouse embryo caused by deficiency in the RNA-editing enzyme ADAR1. J. Biol. Chem. 279, 4894–4902. Herrick-Davis, K., Grinde, E., Niswender, C.M., 1999. Serotonin 5HT2CR receptor RNA editing alters receptor basal activity: implications for serotonergic signal transduction. J. Neurochem. 73, 1711–1717. Higuchi, M., et al., 2000. Point mutation in an AMPA receptor gene rescues lethality in mice deficient in the RNA-editing enzyme ADAR2. Nature 406, 78–81.
Iwamoto, K., Kato, T., 2003. RNA editing of serotonin 2C receptor in human postmortem brains of major mental disorders. Neurosci. Lett. 346, 69–72. Kozak, M., 1980. Influence of mRNA secondary structure on binding and migration of 40S ribosomal subunits. Cell 19, 79–90. Lehmann, K.A., Bass, B.L., 2000. Double-stranded RNA adenosine deaminases ADAR1 and ADAR2 have overlapping specificities. Biochemistry 39, 12875–12884. Lim, V.I., 1995. Analysis of action of the wobble adenine on codon reading within the ribosome. J. Mol. Biol. 252, 277–282. Liu, Y., Emeson, R.B., Samuel, C.E., 1999. Serotonin-2C receptor pre-mRNA editing in rat brain and in vitro by splice site variants of the interferon-inducible double-stranded RNA-specific adenosine deaminase ADAR1. J. Biol. Chem. 274, 18351–18358. Maas, S., Rich, A., Nishikura, K., 2003. A-to-I RNA editing: recent news and residual mysteries. J. Biol. Chem. 78, 1391–1394. Marion, S., Weiner, D.M., Caron, M.G., 2004. RNA editing induces variation in desensitization and trafficking of 5-hydroxytryptamine 2c receptor isoforms. J. Biol. Chem. 279, 2945–2954. McGrew, L., Price, R.D., Hackler, E., Chang, M.S., Sanders-Bush, E., 2004. RNA editing of the human serotonin 5-HT2C receptor disrupts transactivation of the small G-protein RhoA. Mol. Pharmacol. 65, 252–256. Niswender, C.M., Copeland, S.C., Herrick-Davis, K., Emeson, R.B., SandersBush, E., 1999. RNA editing of the human serotonin 5-hydroxytryptamine 2c receptor silences constitutive activity. J. Biol. Chem. 274, 9472–9478. Niswender, C.M., et al., 2001. RNA editing of the human serotonin 5HT2CR receptor: alterations in suicide and implications for serotonergic pharmacotherapy. Neuropsychopharmacology 24, 478–491. Pompeiano, M., Palacios, J.M., Mengod, G., 1994. Distribution of the serotonin 5-HT2 receptor family mRNAs: comparison between 5-HT2A and 5HT2CR receptors. Brain Res. Mol. Brain Res. 23, 163–178. Price, R.D., Sanders-Bush, E., 2000. RNA editing of the human serotonin 5-HT (2C) receptor delays agonist-stimulated calcium release. Mol. Pharmacol. 58, 859–862. Price, R.D., Weiner, D.M., Chang, M.S., Sanders-Bush, E., 2001. RNA editing of the human serotonin 5-HT2C receptor alters receptor-mediated activation of G13 protein. J. Biol. Chem. 276, 44663–44668. Schaefer, M., et al., 2002. Interferon alpha (IFNalpha) and psychiatric syndromes: a review. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 26, 731–746. Seal, S.N., Schmidt, A., Marcus, A., 1989. Ribosome binding to inosine-substituted mRNAs in the absence of ATP and mRNA factors. J. Biol. Chem. 264, 7363–7368. Seeburg, P.H., Hartner, J., 2003. Regulation of ion channel/neurotransmitter receptor function by RNA editing. Curr. Opin. Neurobiol. 13, 279–283. Tecott, L.H., et al., 1995. Eating disorder and epilepsy in mice lacking 5HT2CR serotonin receptors. Nature 374, 542–546. Wang, Q., O'Brien, P.J., Chen, C.-X., Cho, C.D.-S., Murray, J.M., Nishikura, K., 2000. Altered G protein-coupling functions of RNA editing isoform and splicing variant serotonin 2c receptors. J. Neurochem. 74, 1290–1300. Wang, Q., et al., 2004. Stress-induced apoptosis associated with null mutation of ADAR1 RNA editing deaminase gene. J. Biol. Chem. 279, 4952–4961. Yang, W., Wang, Q., Kanes, S.J., Murray, J.M., Nishikura, K., 2004. Altered RNA editing of serotonin 5HT2CR receptor induced by interferon: implications for depression associated with cytokine therapy. Mol. Brain Res. 124, 70–78.
A-to-I pre-mRNA editing of the serotonin 2C receptor: Comparisons among inbred mouse strains Yunzhi Du a,b , Muriel T. Davisson e , Karen Kafadar d , Katheleen Gardiner a,b,c,⁎ a
Eleanor Roosevelt Institute at the University of Denver, USA Program in Human Medical Genetics, University of Colorado at Denver and the Health Sciences Center, Denver, Colorado, USA Department of Biochemistry and Molecular Biology, University of Colorado at Denver and the Health Sciences Center, Denver, Colorado, USA d Department of Mathematics, University of Colorado at Denver and the Health Sciences Center, Denver, Colorado, USA e The Jackson Laboratory, Bar Harbor, Maine, USA b
c
Received 20 April 2006; received in revised form 30 May 2006; accepted 7 June 2006 Available online 27 June 2006
Abstract The serotonin receptor 5HT2CR pre-mRNA is subject to adenosine deamination (RNA editing) at five residues located within a 15 nucleotide stretch of the coding region. Such changes of adenosine to inosine (A-to-I) can produce 32 mRNA variants, encoding 24 different protein isoforms, some of which vary in biochemical and pharmacological properties. Because serotonin mediates diverse neurological processes relevant to behavior and because inbred mouse strains vary in their responses to tests of learning and behavior, we have examined the A-to-I editing patterns of the 5HT2CR mRNA in whole brains from eight mouse strains. By sequencing approximately 100 clones from individual mice, we generated detailed information on levels of editing at each site and patterns of editing that identify a total of 28 mRNA and 20 protein isoforms. Significant differences between individuals from different strains were found in total editing frequency, in the proportion of transcripts with 1 and 4 edited sites, in editing frequency at the A, B, E and D sites, in amino acid frequencies at positions 157 and 161, and in subsets of major protein isoforms. Primer extension assays were used to show that individuals within strains (six C3H.B−+rd1 and four 129SvImrJ) displayed no significant differences in any feature. These findings suggest that genetic background contributes to subtle variation in 5HT2CR mRNA editing patterns which may have consequences for pharmacological treatments and behavioral testing. © 2006 Elsevier B.V. All rights reserved. Keywords: 5HT2C receptor; Inosine; Protein isoforms; Cloning; Primer extension; Behavior
1. Introduction The serotonin receptor 2C (5HT2CR) is widely distributed within the central nervous system (Pompeiano et al., 1994; Liu et al., 1999). It mediates diverse neurological processes that affect feeding behavior, sleep, sexual behavior, anxiety and depression (Tecott et al., 1995; Brennan et al., 1997, reviewed by Graeff et al., 1996). The 5HT2CR pre-mRNA is the only G protein coupled receptor transcript known to be a target of site-specific adenosine deamination (A-to-I pre-mRNA editing). Editing can produce Abbreviations: 5HT2CR, serotonin receptor 2C; I, inosine; A, adenosine; G, guanosine. ⁎ Corresponding author. Eleanor Roosevelt Institute at the University of Denver, 1899 Gaylord Street, Denver, Colorado, 80206, USA. Tel.: +1 303 336 5652; fax: +1 303 333 8423. E-mail address: [email protected] (K. Gardiner). 0378-1119/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2006.06.007
inosine at five sites (A, B, E, C, D) within a 15 nucleotide segment located within the putative second intracellular loop (Burns et al., 1997). These editing sites are conserved between human and rodents, although predominant editing patterns and levels differ (Niswender et al., 1999). Because inosine is interpreted by the translation machinery as guanosine, editing can change the genomically encoded amino acid sequence 157-IRNPI-161 at I, N and I, as shown in Fig. 1. Considering editing independently at each site, 32 mRNA variants encoding 24 protein isoforms are theoretically possible (Burns et al., 1997). Editing of 5HT2CR mRNA affects biochemical and pharmacological properties of the protein. Of the isoforms tested, the unedited INI and the fully edited VGV isoforms display the extremes of several properties, with INI the most active isoform in the absence of agonist, and VGV the least, and INI displaying the highest agonist affinity and potency, and VGV the lowest (Niswender
40
Y. Du et al. / Gene 382 (2006) 39–46
TJL's IACUC. Eight-week old male mice from inbred strains: C57BL/6J, DBA/2J (DBA), C3H.B− +rd1 (C3H ), 129S1/SvImJ (129), FVB/NJ ( FVB), BUB/BnJ (BUB) and CAST/Ei (CAST ), and the F1 hybrid B6C3.b F1 (B6C) (obtained by crossing C57BL/6J and C3H.B− + rd1 ) were sacrificed by cervical dislocation, all between 9 and 11 AM. One mouse was used from each strain, except for two C57BL/6J, seven C3H and five 129. Brains were dissected out and immediately frozen in liquid nitrogen. Fig. 1. Edited region of the 5HT2C receptor. Nucleotides 1159–1173 of the mouse 5HT2CR mRNA (NM_008312) identical to the genomic sequence are shown. Adenosine residues that can be edited are labeled as sites A–E (site E is also known as site C′). The genomically encoded amino acid sequence from positions 157–161, is IRNPI, abbreviated as INI in the completely unedited form. Amino acid changes corresponding to each editing event are indicated.
et al., 1999; Herrick-Davis et al., 1999). Other, partially edited, isoforms typically display intermediate characteristics. The predominant isoform in rodent brain, VNV, shows significantly decreased receptor basal activity compared with INI (Herrick-Davis et al., 1999), and VSV, MSV, IDV, IGV and VGI result in decreased inositol phosphate release in response to serotonin (Wang et al., 2000). Lastly, basal activity of INI is significantly reduced by exposure to the antipsychotic drugs loxapine and clozapine, while VGV is unaffected (Niswender et al., 2001), and, while INI and VGV bind lysergic acid diethylamide (LSD) with similar affinity, only VGV exhibited a markedly reduced response, as measured by phosphoinositide hydrolysis (Niswender et al., 2001). These data suggest that alteration of editing patterns or levels, whether by genetic, environmental or stochastic processes, may have neurological consequences affecting behavior and responses to the environment and pharmacological agents. Indeed, several recent reports have described statistically significant differences in editing patterns between controls and subsets of patients with depression (Gurevich et al., 2002a; Niswender et al., 2001), and between controls and mice treated with a 5HT2C agonist or antagonist (Gurevich et al., 2002a,b), including some mouse strain differences (Englander et al., 2005; Hackler et al., 2006). To further explore the genetic basis of 5HT2CR editing patterns, with a long term goal of uncovering additional correlations between 5HT2CR editing levels and behavior, we have examined editing patterns in whole brains from eight strains of mice. Individuals within an inbred strain are genetically identical, and members of different strains are known to differ in their responses to specific behavioral tests (reviewed in Crawley et al., 1997), including those designed to assess anxiety (Bouwknecht and Paylor, 2002). Comparison of editing pattern differences within and among strains suggests that there are biologically subtle, but statistically significant, genetic effects that may be relevant to 5HT2CR function. 2. Materials and methods 2.1. Mouse brain tissue All mice were raised and sacrificed at The Jackson Laboratory (TJL) under TJL animal protocol number 99092 approved by
2.2. Cloning and analysis of 5HT2CR sequences RNA was extracted from individual brains from each strain, using the standard guanidine-phenol method and RNA integrity was verified by gel electrophoresis. One microgram of total RNA was used in first-strand cDNA synthesis with MMLV reverse transcriptase and random hexamer priming (Invitrogen). The segment spanning the edited region was amplified by the polymerase chain reaction ( PCR) with primers 5′ TGT GCT ATT TTC AAC TGC GTC CAT CAT G 3′ and 5′ CGG CGT AGG ACG TAG ATC GTTAAG 3′ located in exon 5 and exon 6, respectively, of the mouse 5HT2CR mRNA sequence ( NM_008312). Fresh PCR product (327 bp) was cloned into PCRII 2.1 ( Invitrogen) and transformed into E. coli. For each brain, approximately 100 individual clones were picked from a single transformation. Inserts were sequenced on both strands using Big Dye Terminator 3.0 cycle sequencing kit (ABI) and M13 reverse and forward primers 5′ CAGGAAACAGCTATGAC3′ and 5′GTAAAACGACGGC CAG3′. Sequence traces were examined for the presence of an A or a G residue at each of the five editing sites, and the complete sequence of the edited region of each clone was scored. Sequences with high background or ambiguous base calling were discarded. Between 80 and 106 sequences were determined for each brain. 2.3. Primer extension assays Total RNA isolation, reverse transcription and PCR were performed as described above on six additional brains from C3H and four additional brains from 129 (these did not include samples used in cloning experiments). PCR products were purified through ChromaSpin TE100 columns (Clontech). Forty ng of PCR product was used as template for primer extension assays. Primers and assay conditions were modified from Chen et al. (2000) and Niswender et al. (2001). Forward strand primer extA18mer, 5′CGCTGGACCGGTATGTAG3′, was used to measure A site editing frequency in the presence of 4.3 mM ddTTP and ddGTP, 0.7 mM dATP and dCTP. Forward strand primer extA20mer, 5′CGCTGGACCGGTATGTAGCA3′, was used to measure B site editing frequency in the presence of 6 mM ddATP and 1 mM each dTTP, dGTP and dCTP or 6 mM ddGTP, and 1 mM dATP, dTTP and dCTP. Reverse strand primer extD, 5′ GAATTGAACCGGCTATGCTC3′, was used to measure D, C and E site editing frequencies in the presence of 6 mM ddTTP, 1 mM dATP, dGTP and dCTP or 6 mM ddCTP, 1 mM dATP, dTTP and dGTP. Annealing was carried out at 25 °C for 30 min, and extension, at 37 °C, for 10 min in the presence of 7 mM
Y. Du et al. / Gene 382 (2006) 39–46 Table 1 Site-specific editing frequencies
41
and ethanol precipitation. The region putatively involved in RNA duplex formation, containing the edited sites in exon 5 and the 5′ end of intron 5, was amplified by PCR using primers 5′ GTATGGTCAGTGCAATTTGGATAAC3′ and 5′GTAT CAGTGTTGCCAAAATCCACTG3′, located in introns 4 and 5, respectively, of the mouse genomic sequence ( NC_000086). The PCR product (518 bp) was sequenced directly using the same primers, as described above. 3. Results 3.1. Cloning and sequencing
Percentage of clones with a G residue at each of the sites A–E was calculated from the number of clones sequenced. Significant differences ( p < 0.05) are indicated in bold (high) and bold shaded (low); * trend towards significance (0.05 < p < 0.1). P values are indicated in the text. For comparison with primer extension data in Table 7, 95% confidence limits for A, B C and D sites are between 6% and 11%.
MnCl2, 10 mM DTT, 1× sequencing buffer, 1.5 units Sequenase version 2.0 (USB). Products were resolved on standard denaturing 15% polyacrylamide gels, exposed to a phosphor screen and scanned on a PhosphoImager (Molecular Dynamics, Strom 840). Signal intensities of bands and relative ratios were quantified using ImageQuant (Molecular Dynamics).
Between 81 and 106 clones per individual, for a total 740 clones, were sequenced from one individual from each of seven inbred strains and one F1 hybrid. An additional 29 clones were sequenced from a second individual from C57BL/6J. Each sequence was scored for the presence of a G residue, indicating editing, at each of the five sites, A–E, shown in Fig. 1. Sequences were used to calculate the following editing patterns in individual mice: the proportion of editing at individual sites, the proportion of editing at all five sites, the proportion of transcripts edited at different numbers of sites, the proportion of each mRNA isoform, the occurrence of specific amino acids at individual sites, and the occurrence of each protein isoform. Data for all mice are summarized in Tables 1–6. Pairwise comparisons were carried out between individual mice.
2.4. Statistical analyses 3.2. Variation in editing frequency at individual sites All statistical analysis used the GraphPad Prism4 software. Differences in editing frequency at five editing sites and in amino acid frequencies and protein isoforms obtained from sequence analysis of clones from individual mice were analyzed using Fisher's Exact Test to compare all possible pairs. The Modified Wold method was used to determine 95% confidence limits in individual site editing levels. Results of quantification of primer extension assays were analyzed with Two-Tailed Unpaired Students t-test. 2.5. Analysis of 5HT2CR genomic sequence Genomic DNA was isolated from the livers of each mouse analyzed in the cloning experiment using standard proteinase K digestion plus phenol chloroform extraction, RNase A digestion
Table 2 Percentage of transcripts with 0–5 inosines
Bold, shading and * as in Table 1.
As shown in Table 1, editing levels at the A site in C3H, BUB and B6C, at 89%, 87% and 81%, respectively, were significantly higher than those in DBA and 129, at 67% and 65% ( p < 0.001, p < 0.01 and p < 0.05). C3H and BUB were also significantly higher than FVB ( p < 0.01 and 0.05). Editing at the B site was significantly higher in BUB, C3H and B6C (73%–82%) relative to 129 (52%) ( p < 0.001, p < 0.0001 and p < 0.05, respectively). C3H was also higher than DBA ( p < 0.01). At the E site, C3H and BUB showed significantly higher levels (10% and 9%) than FVB (1%) ( p < 0.05) and a trend toward higher levels than CAST at 2% ( p = 0.06). No significant differences were found for the editing frequency at the C site. At the D site, B6C editing, at 79%, was significantly higher than CAST at 66% ( p < 0.05); C3H (79%)
Table 3 Frequencies of major mRNAs and unedited and completely edited variants
Bold, shading as in Table 1.
42
Y. Du et al. / Gene 382 (2006) 39–46
Table 4 Editing at A and B sites is not independent
*, + sign, denotes editing to an inosine residue; − sign, denotes an unedited, genomic A Residue. Subscripts refer to the editing status of the context site, i.e. A+B+ is the ratio of A+B+ to (A+B+ + A−B+), where edited B site is the context site. Bold, indicates all values in the column are significantly greater than values in the adjacent column in bold shaded.
showed a trend toward a significant difference from CAST ( p < 0.07). The last column of Table 1 shows the total amount of editing at all five sites. C3H, at 57%, had significantly higher total editing than DBA, 129, FVB, CAST and C57BL/6J at 44%– 49% ( p < 0.05 to p < 0.0001). BUB and B6C, at 52%, had higher levels than 129 (44%, p < 0.05). Cloning and sequencing were carried out for a smaller number of clones (29) from a second brain of C57BL/6J. Editing levels were not significantly different between the two mice (Table 1). 3.3. Variation in editing levels among transcripts The presence of inosine within a transcript may have consequences for translational efficiency, especially when there are multiple inosines within a short stretch. Table 2 shows the proportion of transcripts edited at any combination of 1 to 4 sites. C3H has the highest proportion of transcripts, 26%, edited at four sites, and the lowest proportion of transcripts, 7%, edited at a single site. 129, FVB and CAST show the opposite proportions, with only 10%–12% edited at four sites but 20%–22% edited at a single site. These differences are significant ( p < 0.05). Other significant differences in the level of transcripts with one site edited include C3H < DBA, BUB < 129, and B6C < 129 and CAST ( p < 0.05). C3H and 129 also show trends towards significant differences in the proportion of transcripts edited at three sites and two sites, with 50% and 13%, respectively in C3H, compared with 37% and 23%, in 129 ( p = 0.09 and p = 0.08, respectively). 3.4. Variation in mRNA editing patterns Differences in mRNA editing patterns may exist that are not apparent from editing frequencies at individual sites. In this analysis, a clone sequence labeled ABD indicates sites A, B and D were edited to inosines, while sites C and E remained the genomic A residues. Such combinations of editing theoretically can produce 32 different RNA sequences. The frequencies of the six most commonly observed variants are shown in Table 3. The frequencies of the fully edited and completely unedited forms are included for comparison. In all eight strains, the predominant form was ABD, ranging from 33% to 46% and
representing an average of 37% of mRNAs in all samples. The other common mRNA variants were ABCD with an average frequency of 12%, AB with 8%, D only with 10%, and AD, A only and the completely unedited form, each averaging ∼ 6%. The fully edited form was very rare, seen only in two mice and only at 1%. Four mRNA variants were not observed in the entire 740 clone set: BE, BEC, BED and BECD. The remaining 20 variants were observed in fewer than 5% of clones in each mouse. In comparing individual mice, the D site variant occurred less frequently in C3H (2%) than in 129, FVB, DBA and CAST (12%–17%, p < 0.01–0.05), and in BUB less often (4%) than in 129 ( p < 0.01). 3.5. Independence of editing at adjacent sites Enzymes that carry out editing, Adenosine Deaminases that Act on RNA, ADAR1 and ADAR2, display preferences in nearest neighbors surrounding editing sites (Lehmann and Bass, 2000; Dawson et al., 2004). Studies with mouse models with null mutations in one or both enzymes suggest that in 5HT2CR the A and B sites are predominantly edited by ADAR1 and the E, C and D sites, by ADAR2 (Higuchi et al., 2000; Wang et al., 2004; Hartner et al., 2004). Because the five 5HT2CR sites are so closely spaced, we asked if editing at one site influenced the observed frequency of editing at adjacent sites. For this, we calculated the percent editing at each site in the context of +/− editing at the immediately adjacent site. For example, A+B+ is
Table 5 Frequencies of cDNAs encoding major protein isoforms and V at 157
Bold, shading and * as in Table 1.
Y. Du et al. / Gene 382 (2006) 39–46
43
Table 6 Frequencies of cDNAs encoding minor protein isoforms
C57BL/6J DBA C3H 129 FVB BUB CAST B6C
ISI
ISV
IDI
IDV
IGI
IGV
VDI
VDV
VGI
VGV
MNI
MNV
MSI
MSV
Total
0% 2% 0% 1% 1% 0% 1% 0%
3% 3% 1% 5% 2% 0% 2% 2%
0% 0% 1% 0% 0% 1% 0% 0%
2% 0% 0% 1% 0% 2% 0% 0%
0% 0% 0% 1% 0% 0% 0% 0%
1% 0% 0% 0% 0% 0% 0% 1%
1% 0% 1% 0% 0% 2% 1% 0%
0% 2% 4% 2% 1% 3% 1% 6%
1% 1% 1% 1% 0% 0% 0% 1%
1% 0% 2% 0% 0% 0% 0% 1%
0% 1% 0% 0% 1% 0% 0% 0%
0% 3% 1% 1% 4% 1% 2% 1%
0% 1% 0% 0% 0% 0% 0% 0%
0% 0% 1% 0% 0% 0% 0% 0%
9% 13% 12% 12% 9% 9% 7% 12%
the percent A sites edited in the context of B site edited, i.e. the ratio of A+B+ to (A+B+ + A−B+), where + denotes an I residue and − denotes a genomic A residue. Similarly, A+B− is the ratio of A+B− to (A+B− + A−B−). Results in Table 4 assess the independence of editing at sites A and B, E and C, and C and D. Editing at the A and B sites (which are separated by a single U residue) is not independent; if either A or B is edited, it is highly likely that the second site is also edited (p < 0.0001, paired Student t-test); this may in part be a reflection of both sites being edited by ADAR1. In contrast, editing at the E site (immediately 5′ to the C site) is very low regardless of C site status, and editing events at the C and D sites (separated by four nucleotides) are independent of each other. 3.6. Prediction of consequences for protein isoforms Editing at any of the five sites alone causes a codon change as indicated in Fig. 1. It is the protein coding consequences that will provide information on the functional implications of 5HT2CR mRNA editing. While 28 mRNA isoforms were observed, these coded for 20 of the 24 possible protein isoforms. Table 5 shows the frequencies of cDNAs encoding the six most common protein isoforms found and Table 6, the frequencies of the 14 most rarely observed cDNAs. These data allow direct extrapolation to the 5HT2CR isoform population, if it is assumed that all cDNA variants are transported and translated with uniform efficiency. Editing at the A site, +/− editing at the B site, changes the genomically-encoded isoleucine (I) to valine (V), and therefore the significant differences in A site editing noted above would be reflected in the frequency of protein isoforms containing V at position 157. At position 159, C3H and CAST had the highest percentage of clones (31%) that had been edited to encode S, D or G, however, these are not significantly different from the lowest (21%) seen in 129. Based on differences in editing at the D site, a significant difference in the presence of V at position 161 is predicted in comparing B6C with CAST, and a trend towards significance in comparing C3H and CAST (Table 1). At the individual isoform level, the variation in the frequency of D only isoform, noted above, would result in the same variations among strains of the frequency of the INV isoform (Table 5). Significantly higher levels of VSI (ABC and AC) are predicted in CAST (12%) relative to C3H (2%) ( p < 0.05), 129 (1%) ( p < 0.01) and B6C (3%) ( p < 0.05). Fourteen cDNA variants that encode minor protein isoforms are observed on average at < 1% in all strains. These are the
forms that contain methionine (M), and a subset of those containing glycine (G) and aspartic acid (D), and result from editing at B but not A, editing at E plus C, and editing at E but not C, respectively (Fig. 1). Table 6 lists the frequencies of these isoforms in the eight brains; typically fewer than half are represented in any sample, in total representing less than 13% of all isoforms. The four forms encoding Met plus Gly or Asp (MGV/I and MDV/I) were never observed. Lastly, although functional information is incomplete, subsets of isoforms can be grouped based on some biochemical properties. Herrick-Davis et al. (1999) measured basal activities of seven isoforms and ranked them as follows: if INI is assigned the maximal basal activity of 100%, ISV, INV and VSI display ∼ 75%, VNV and VSV, ∼ 50% and VGV, ∼ 22% of the maximal. The eight strains do not differ in levels of INI or VGV, but when isoforms of intermediate activities are grouped, several significant differences are seen (Table 5, columns 8,9). In particular, C3H has the highest levels (67%) of the less active VNV/VSV group and the lowest levels (5%) of the more active INV/ISV/VSI group. In contrast, DBA, 129 and CAST all have both lower levels (48%–50%) of the less active and higher levels (19%–26%) of the more active isoforms. Based on these data, C3H would be predicted to have overall less basal 5HT2CR activity compared with DBA, 129 and CAST (p < 0.05). In considering only the more active group, 129 has higher levels than BUB; CAST has higher levels than BUB and B6C; and FVB and C57BL/6J have higher levels than C3H. 3.7. Editing frequencies in C3H and 129 individuals Because all individuals from an inbred mouse strain are genetically identical to each other, if 5HT2CR editing patterns are solely genetically determined, no significant differences are expected among individuals within a strain. To examine individual variation, primer extension assays were used to measure editing frequencies in six additional individuals from C3H and four additional individuals from 129. These strains were chosen because they displayed the greatest differences in results from cloning and sequencing experiments. Representative gels are shown in Fig. 2 and results are summarized in Table 7. No significant differences in editing levels are seen within either strain, i.e. either among the six C3H individuals or among the four 129 individuals. In addition, at the A and B sites in 129 and at the C and D sites in both strains, levels of editing fall
44
Y. Du et al. / Gene 382 (2006) 39–46
Fig. 2. Representative primer extension gels. Primer extension assays were carried out on cDNA amplified from total brain RNA prepared from C3H and 129 individuals (#1 and #2), as described in the Materials and methods. (a) primer extA-18mer with ddGTP plus ddTTP, (b) primer extA-20mer with ddATP and ddGTP reactions (A and G, respectively), (c) primer extD with ddTTP and ddCTP reactions (T and C, respectively). Bands are identified by editing status (+, edited; −, unedited) and site (A, B, D, C or E); sizes of bands are indicated in nucleotides. P, primer.
within the 95% confidence limits of the cloning results shown in Table 1 (levels of editing at the A and B sites have been reported to be artefactually low by this method (Burns et al., 1997) and this may account for the small discrepancy in C3H from the values obtained by cloning). Importantly, editing at the A and B sites in C3H at 78% and 65%, respectively, is significantly higher than that in 129 at 70% and 54%, consistent with the cloning data. 3.8. Sequence of exon 5/intron 5 A-to-I pre-mRNA editing requires a double stranded RNA structure as a substrate (Bass, 2002; Maas et al., 2003; Seeburg and Hartner, 2003). In 5HT2CR pre-mRNA, this is an imperfect helix formed between the segment of exon 5 containing the edited region and the 5′ end of the downstream intron 5 (Burns et al., 1997). Variations in editing levels could be caused by polymorphisms that alter the stability of the double stranded structure so that enzyme binding is perturbed. To address this
Table 7 Editing frequency of individuals from C3H and 129 strains
Measured by primer extension as in Fig. 2.
possibility, the genomic sequence of exon five and the 5′ end of intron 5 encompassing the double stranded region was determined. All mice were homozygous in this region, and the nucleotide sequences were identical to previously published sequences (acc# NC_000086; data not shown). 4. Discussion In designing these experiments, we decided to examine individual mice and to use two approaches to determining 5HT2CR editing patterns. First, we used the more laborious, but more informative, approach of cloning cDNAs spanning the 5HT2C receptor editing site region and sequencing a large number of clones. This approach allowed us to determine frequencies of complete protein isoforms, to observe more rare protein isoforms, and to achieve statistical significance in several observed differences. Sequencing of a smaller number of clones from a second individual of one strain showed these two individuals did not differ significantly in editing levels and patterns. We then used the more rapid primer extension method to measure editing patterns of additional individuals from two different strains. Primer extension yields less detailed information on editing patterns and does not provide complete isoform information. However, statistical analysis of these data also indicated that individuals within a single strain did not differ significantly in their editing levels. This suggests the predominance of genetic regulation of editing over potential environmental and stochastic events. Genomic sequencing of the region believed to be involved in editing site selection showed no differences among strains, and therefore other mechanisms, such as activity levels of ADAR1 or ADAR2 that carry out editing, may be responsible for the inter-strain differences in editing levels. Based on these data, however, we conclude that it is reasonable to use the data from the cloning experiments in further strain comparisons. Sequencing 740 clones from the eight individual mouse brains identified a total of 28 of a possible 32 mRNA variants, encoding 20 of a possible 24 protein isoforms. This describes the most complete repertoire of isoforms so far observed in
Y. Du et al. / Gene 382 (2006) 39–46
mammalian brain. Only six isoforms were common and seen in all individuals (Table 5). The remaining 14 isoforms were rare (Table 6); only 5–8 of these were seen in any one brain and several were seen only in 1%–2% of sequences in only one or two brains. These general results are consistent with previous analyses in rodent brain (Burns et al., 1997; Niswender et al., 1999). Statistically significant differences in the frequency of editing at the A, B, E and D sites, in the total amount of editing, in the frequency of protein isoforms with V at position 157 and 161, and in the INV and VSI isoforms were observed in subsets of pairwise comparisons between strains. While whole brains were used here, it is known that editing levels vary, sometimes dramatically, among brain regions in humans and rodents (Niswender et al., 2001; Wang et al., 2000; Burns et al., 1997). Therefore, it is possible that some isoforms may be much more common and/or that variations between strains may be different when separate brain regions are examined, for example as has been shown in amygdala from C57BL/6J compared with DBA (Hackler et al., 2006). Biochemical and pharmacological consequences of editing have been assessed for some of the 24 isoforms. Most data have been generated for the completely unedited isoform, INI, the completely edited form, VGV, and occasionally for some partially edited forms (Berg et al., 2001; Price et al., 2001; Price and Sanders-Bush, 2000; Fitzgerald et al., 1999; McGrew et al., 2004; Marion et al., 2004). INI and VGV are too rare in whole brain to evaluate variation in their levels. However, for INV, ISV and VSI, Herrick-Davis et al. (1999) showed that basal activities were reduced by approximately 25% compared with INI. This is a less dramatic reduction than the approximately 50% seen with VNV and VSV. The proportions of isoforms with these relatively high and low basal activities are significantly different in C3H, and predict that C3H will have overall low basal activity relative to 129, DBA and CAST (Table 5). It would be useful to extend comprehensive isoform-specific comparisons to other biochemical and pharmacological properties, including coupling to different G proteins, activation of phospholipase A2 (PLA2) versus phospholipase C (PLC) for production of arachidonic acid and phosphoinositol, and responses to therapeutic and recreational drugs. If the INV/ISV/VSI isoform group typically differs from the VSV/VNV group in the strength of functional properties, these strains may differ in additional specific features of serotonin signaling. It is also of interest to consider protein levels of edited isoforms. Indirect evidence suggests that inosine in mRNA may destabilize translation through altered secondary structures or altered base pairing efficiency of tRNAs with inosine-containing codons. Most studies examined inosine in the 5′ UTR of mRNA or in the wobble position of tRNA (Kozak, 1980; Seal et al., 1989; Lim, 1995). Consequences may be significant in coding regions for first and second codon positions which occur in the A, E, C and D sites of 5HT2CR. Total editing levels of 5HT2CR vary among strains, with C3H displaying the highest levels, at 57%, compared to DBA, 129 and CAST (Table 1). In addition, however, C3H has the lowest proportion of transcripts with one inosine and the highest with four inosines (Table 2). If four inosines are relatively inefficiently translated, C3H may
45
have an isoform profile at the protein level more similar to 129, DBA and CAST, and/or may have lower total 5HT2CR protein levels. Editing is not a static feature. Yang et al. (2004) showed that treatment of human glioblastoma cell lines with Interferon-α resulted in dramatic increases in editing at the A, B and C sites (from <40% to >65%), and in the proportion of VSI isoform (from <10% to >50%). Interferon-α is used in treatment of some malignancies and has been associated with onset of depression (Schaefer et al., 2002). Gurevich et al. (2002a,b) showed that treatment of mice with fluoxetine (Prozac), a competitive antagonist of 5HT2CR, decreased editing at the E site from 9% to 1% and at the C site from 30% to 20%. Depletion of 5HT by treatment with an irreversible inhibitor of tryptophan hydroxylase produced similar results. Consistent with these data, treatment with the 5HT2CR agonist DOI increased editing at the E site. The authors concluded that editing levels are being modulated to regulate sensitivity to 5HT as environmental factors change. When initial levels of editing differ among individuals, however, it is necessary to determine if drug treatments, or environmental factors, maintain, abolish or otherwise modulate such differences. This is an important consideration for human treatments, where it has been shown that levels of editing, both site and isoform specific, vary considerably among normal controls (Gurevich et al., 2002a; Dracheva et al., 2003; Iwamoto and Kato, 2003). Comparisons among inbred strains, such as 129S1/SvImJ, FVB/NJ and C3H.B− +rd1, that normally differ in editing features affected by interferon and fluoxetine, will be useful for these experiments. However, it will also be necessary to consider that the predominant isoforms seen in human brain (e.g. VSV, INI and VNV at 38%, 10% and 8%) (Niswender et al., 2001; see also Fitzgerald et al., 1999) differ from those in rodent brain shown in Table 4 (see also Burns et al., 1997; Wang et al., 2000). Differences in editing patterns may also be relevant to responses to stressful environments. For example, strains 129 and C3H differ in responses to some tests associated with anxiety and also differ in levels of some significant isoforms. In context fear conditioning and the acoustic startle response, 129 performance is described as good and poor, respectively, and C3H as poor and moderate, while both are described as performing poorly in the open field locomotion test (Crawley et al., 1997). Bouwknecht and Paylor (2002) compared nine strains in stress-induced hyperthermia and exploration in the light–dark paradigm, tests which assay physiological and behavioral responses to stress, respectively. Interestingly, behavioral responses did not correlate with physiological responses, implicating multiple mechanisms. 5HT2CR editing patterns certainly would not be the sole factor governing stress or drug responses, but if they change with behavioral tests, particularly in specific cell types and/or brain regions, it will be important to determine if correlations exist between naïve editing patterns, and changes induced by stress and performance in aversive learning and memory tests. Acknowledgements This work was supported by grants from the Cullpepper Foundation and the National Institutes of Health to KG. The
46
Y. Du et al. / Gene 382 (2006) 39–46
authors thank Cecilia Schmidt and Dobromir Slavov for technical assistance. References Bass, B.L., 2002. RNA editing by adenosine deaminases that act on RNA. Ann. Rev. Biochem. 71, 817–846. Berg, K.A., Cropper, J.D., Niswender, C.M., Sanders-Bush, E., Emeson, R.B., Clarke, W.P., 2001. RNA-editing of the 5-HT(2C) receptor alters agonist– receptor–effector coupling specificity. Br. J. Pharmacol. 134, 386–392. Bouwknecht, J.A., Paylor, R., 2002. Behavioral and physiological mouse assays for anxiety: a survey in nine mouse strains. Behav. Brain Res. 136, 489–501. Brennan, T.J., Seeley, W.W., Kilgard, M., Schreiner, C.E., Tecott, L.H., 1997. Sound-induced seizures in serotonin 5HT2CR receptor mutant mice. Nat. Genet. 16, 387–390. Burns, C.M., et al., 1997. Regulation of serotonin-2C receptor G-protein coupling by RNA editing. Nature 387, 303–308. Chen, C.X., Cho, D.S., Wang, Q., Lai, F., Carter, K.C., Nishikura, K., 2000. A third member of the RNA-specific adenosine deaminase gene family, ADAR3, contains both single- and double-stranded RNA binding domains. RNA 6, 755–767. Crawley, J.N., et al., 1997. Behavioral phenotypes of inbred mouse strains: implications and recommendations for molecular studies. Psychopharmacology 132, 107–124. Dawson, T.R., Sansam, C.L., Emeson, R.B., 2004. Structure and sequence determinants required for the RNA editing of ADAR2 substrates. J. Biol. Chem. 279, 4941–4951. Dracheva, S., Elhakem, S.L., Marcus, S.M., Siever, L.J., McGurk, S.R., Haroutunian, V., 2003. RNA editing and alternative splicing of human serotonin 2c receptor in schizophrenia. J. Neurochem. 87, 1402–1412. Englander, M.T., Dulawa, S.C., Bhansali, P., Schmauss, C., 2005. How stress and fluoxetine modulate serotonin 2C receptor pre-mRNA editing. J. Neurosci. 25, 648–651. Fitzgerald, L.W., et al., 1999. Messenger RNA editing of the human serotonin 5HT2C receptor. Neuropsychopharmacology 21, 82S–90S. Graeff, F.G., Guimaraes, F.S., Andrade, T.G.C.S.D., Deakin, J.F.W., 1996. Role of 5HT in stress, anxiety, and depression. Pharmacol. Biochem. Behav. 54, 129–141. Gurevich, I., Tamir, H., Arango, V., Dwork, A.J., Mann, J.J., Schmauss, C., 2002a. Altered editing of serotonin 2c receptor pre-mRNA in the prefrontal cortex of depressed suicide victims. Neuron 34, 349–366. Gurevich, I., Englander, M.T., Adlersberg, M., Siegal, N.B., Schmauss, C., 2002b. Modulation of serotonin 2C receptor editing by sustained changes in serotonergic neurotransmission. J. Neurosci. 22, 10529–10532. Hackler, E.A., Airey, D.C., Shannon, C.C., Sodhi, M.S., Sanders-Bush, E., 2006. 5-HT(2C) receptor RNA editing in the amygdala of C57BL/6J, DBA/2J, and BALB/cJ mice. Neurosci. Res. 55, (Electronic publication ahead of print). Hartner, J.C., Schmittwolf, C., Kispert, A., Muller, A.M., Higuchi, M., Seeburg, P.H., 2004. Liver disintegration in the mouse embryo caused by deficiency in the RNA-editing enzyme ADAR1. J. Biol. Chem. 279, 4894–4902. Herrick-Davis, K., Grinde, E., Niswender, C.M., 1999. Serotonin 5HT2CR receptor RNA editing alters receptor basal activity: implications for serotonergic signal transduction. J. Neurochem. 73, 1711–1717. Higuchi, M., et al., 2000. Point mutation in an AMPA receptor gene rescues lethality in mice deficient in the RNA-editing enzyme ADAR2. Nature 406, 78–81.
Iwamoto, K., Kato, T., 2003. RNA editing of serotonin 2C receptor in human postmortem brains of major mental disorders. Neurosci. Lett. 346, 69–72. Kozak, M., 1980. Influence of mRNA secondary structure on binding and migration of 40S ribosomal subunits. Cell 19, 79–90. Lehmann, K.A., Bass, B.L., 2000. Double-stranded RNA adenosine deaminases ADAR1 and ADAR2 have overlapping specificities. Biochemistry 39, 12875–12884. Lim, V.I., 1995. Analysis of action of the wobble adenine on codon reading within the ribosome. J. Mol. Biol. 252, 277–282. Liu, Y., Emeson, R.B., Samuel, C.E., 1999. Serotonin-2C receptor pre-mRNA editing in rat brain and in vitro by splice site variants of the interferon-inducible double-stranded RNA-specific adenosine deaminase ADAR1. J. Biol. Chem. 274, 18351–18358. Maas, S., Rich, A., Nishikura, K., 2003. A-to-I RNA editing: recent news and residual mysteries. J. Biol. Chem. 78, 1391–1394. Marion, S., Weiner, D.M., Caron, M.G., 2004. RNA editing induces variation in desensitization and trafficking of 5-hydroxytryptamine 2c receptor isoforms. J. Biol. Chem. 279, 2945–2954. McGrew, L., Price, R.D., Hackler, E., Chang, M.S., Sanders-Bush, E., 2004. RNA editing of the human serotonin 5-HT2C receptor disrupts transactivation of the small G-protein RhoA. Mol. Pharmacol. 65, 252–256. Niswender, C.M., Copeland, S.C., Herrick-Davis, K., Emeson, R.B., SandersBush, E., 1999. RNA editing of the human serotonin 5-hydroxytryptamine 2c receptor silences constitutive activity. J. Biol. Chem. 274, 9472–9478. Niswender, C.M., et al., 2001. RNA editing of the human serotonin 5HT2CR receptor: alterations in suicide and implications for serotonergic pharmacotherapy. Neuropsychopharmacology 24, 478–491. Pompeiano, M., Palacios, J.M., Mengod, G., 1994. Distribution of the serotonin 5-HT2 receptor family mRNAs: comparison between 5-HT2A and 5HT2CR receptors. Brain Res. Mol. Brain Res. 23, 163–178. Price, R.D., Sanders-Bush, E., 2000. RNA editing of the human serotonin 5-HT (2C) receptor delays agonist-stimulated calcium release. Mol. Pharmacol. 58, 859–862. Price, R.D., Weiner, D.M., Chang, M.S., Sanders-Bush, E., 2001. RNA editing of the human serotonin 5-HT2C receptor alters receptor-mediated activation of G13 protein. J. Biol. Chem. 276, 44663–44668. Schaefer, M., et al., 2002. Interferon alpha (IFNalpha) and psychiatric syndromes: a review. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 26, 731–746. Seal, S.N., Schmidt, A., Marcus, A., 1989. Ribosome binding to inosine-substituted mRNAs in the absence of ATP and mRNA factors. J. Biol. Chem. 264, 7363–7368. Seeburg, P.H., Hartner, J., 2003. Regulation of ion channel/neurotransmitter receptor function by RNA editing. Curr. Opin. Neurobiol. 13, 279–283. Tecott, L.H., et al., 1995. Eating disorder and epilepsy in mice lacking 5HT2CR serotonin receptors. Nature 374, 542–546. Wang, Q., O'Brien, P.J., Chen, C.-X., Cho, C.D.-S., Murray, J.M., Nishikura, K., 2000. Altered G protein-coupling functions of RNA editing isoform and splicing variant serotonin 2c receptors. J. Neurochem. 74, 1290–1300. Wang, Q., et al., 2004. Stress-induced apoptosis associated with null mutation of ADAR1 RNA editing deaminase gene. J. Biol. Chem. 279, 4952–4961. Yang, W., Wang, Q., Kanes, S.J., Murray, J.M., Nishikura, K., 2004. Altered RNA editing of serotonin 5HT2CR receptor induced by interferon: implications for depression associated with cytokine therapy. Mol. Brain Res. 124, 70–78.