Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Inbred strains of mice have been used to establish genetic contributions to various ethanol-induced behaviors including hypothermia, ataxia, hypnotic responses, and withdrawal (for review see Phillips and Crabbe, 1991). Ethanol-induced ataxia is clinically important and a good measure of intoxication. Several studies using various measures of ataxia have shown differences in initial sensitivity and acute tolerance across inbred strains (Gallaher et al., 1996; Browman and Crabbe, 2000; Deitrich et al., 2000; Gehle and Erwin, 2000; Kirstein and Tabakoff, 2001). The dowel test for ataxia is a fairly simple test of coordination and balance that allows for the measurement of both initial sensitivity and acute tolerance to ethanol. Using nine inbred strains of mice, we found a significant effect of strain on initial sensitivity and acute functional tolerance (AFT) using the dowel test for ataxia, but there was no correlation between these two measures (Kirstein and Tabakoff, 2001).

Alcoholics have been shown to differ in their platelet adenylyl cyclase activity (Tabakoff et al., 1988). Furthermore, individuals who are family-history-positive for alcoholism have been shown to differ in their platelet adenylyl cyclase activity compared with family-history-negative individuals (Menninger et al., 1998). A role for cAMP signaling in neuronal sensitivity to ethanol has been demonstrated by showing that ethanol-induced inhibition of cerebellar Purkinje neurons involves cAMP signaling (Freund and Palmer, 1997). Cyclic AMP signaling has also been implicated more directly in behavioral sensitivity to ethanol through mutational studies in Drosophila. Mutants (e.g., cheapdate) have been identified that have deficits in cAMP signaling and are more sensitive to ethanol on a measure of postural control (Moore et al., 1998). Additionally, studies in mice have demonstrated that modulation of the cAMP signaling pathway can lead to alterations in sensitivity to ethanol-induced ataxia (Durcan et al., 1991; Dar, 1997; Yoshimura et al., 1998).

Recombinant inbred (RI) mice can be used for quantitative trait loci (QTLs) mapping to provisionally identify the location of genes that influence the complex behaviors being studied (for reviews see Gora-Maslak et al., 1991; Plomin et al., 1991; Moore and Nagle, 2000; Wehner et al., 2001). BXD RI mice are derived from C57BL/6J (C57) and DBA/2J (DBA) mice and have been used extensively in addiction research and to map QTLs responsible for alcohol-related phenotypes (Crabbe et al., 1999; Plomin and McClearn, 1993; Crabbe, 2001). In our previous study, the C57 and DBA mice showed significant differences in cAMP accumulation in the cerebellum and in initial sensitivity on the dowel test for ataxia (Kirstein and Tabakoff, 2001). Therefore, the BXD RI strains of mice were expected to show a great deal of genetic variation on these measures. The purpose of the current study was to better locate and characterize the genetic determinants that influence sensitivity to ethanol and cAMP signaling.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects. Male BXD RI mice and the parental strains, C57BL/6J (C57) and DBA/2J (DBA), were purchased from The Jackson Laboratory (Bar Harbor, ME). A total of 238 mice were used. Prior to our experiments, mice were group-housed (one to five littermates per cage) in the University of Colorado Health Sciences Center Animal Research Center and allowed to acclimate to their surroundings for at least 1 week before testing. Food and water were available ad libitum, and mice were maintained on a 12-h light/dark cycle (lights on at 7:00 AM). Behavioral testing was performed in mice of 55 to 80 days of age. Testing was performed between 9:00 AM and 3:00 PM. The Institutional Animal Care and Use Committee of the University of Colorado Health Sciences Center approved all testing procedures.

Dowel Test for Ataxia. Initial sensitivity was measured using the dowel test for ataxia (Erwin and Deitrich, 1996). We considered that the determination of the initial concentration of ethanol close to its site of action in brain (BEC₀) at the time that the animal showed a particular response to ethanol (i.e., loss of balance on the dowel) was a reflection of "initial sensitivity". Initial sensitivity was, therefore, defined by the blood ethanol concentration (BEC₀; blood derived from the retro-orbital sinus) at loss of balance on the dowel after the first ethanol injection. Prior to ethanol injection, mice were trained to balance on the stationary wood dowel [0.5-inch diameter (1.27 cm)] 50 cm above a floor of wood shavings until a criterion of 5 min was met. Most mice met this criterion during the first try; those that did not met this criterion upon the second attempt. Mice were then given an injection of ethanol [1.75 g/kg, 10% (w/v) in saline i.p.] and placed on the dowel. Once mice lost their balance and fell from the dowel, a blood sample was obtained from the retro-orbital sinus, and this was designated BEC₀. The mice were then tested intermittently for their ability to remain on the dowel for 1 min. At this time, another blood sample was obtained and designated BEC₁. The mice then received a second injection of ethanol [2 g/kg, 10% (w/v) i.p.]. Mice were tested at 5-min intervals until they regained their balance on the dowel, and upon regain of balance, a third blood sample was obtained (BEC₂). Acute functional tolerance (AFT) was defined as the difference between BEC₂ and BEC₁ (AFT = BEC₂  BEC₁). Blood ethanol concentrations were determined by enzymatic assay (Lundquist, 1959). Times for loss and regain of balance were recorded for all mice. A group of 10 to 12 mice were tested for initial sensitivity and acute functional tolerance each day. Depending on the availability of specific strains, mice were tested within the designated age range, and mice of three to five strains were tested per day. C57 and DBA mice were included as subjects on each test day throughout the study.

Whole-Cell Assay of cAMP Production. Cyclic AMP accumulation was measured in whole-cell preparations of cerebellum using the method of Shimizu et al. (1969) with modifications as reported by Kirstein and Tabakoff (2001). The cerebellum was dissected on ice from the brains of mice exactly 1 week after behavioral testing. Briefly, brains were removed rapidly from decapitated mice and rinsed with ice-cold Krebs-Ringer bicarbonate (KRB) buffer on an ice-cold dissection block. The cerebellum was obtained by dissection from the brain stem and was placed in ice-cold KRB buffer containing 120 mM NaCl, 5 mM KCl, 1.3 mM CaCl₂, 1.2 mM MgCl₂, 1.0 mM KH₂PO₄, 20 mM NaHCO₃, and 11.1 mM glucose that was gassed with 95% O₂/5% CO₂. All chemicals were obtained from Sigma-Aldrich (St. Louis, MO), unless noted otherwise. Brain tissue was minced using a McIlwain tissue chopper (The Mickle Laboratory Engineering Co. Ltd., Guildford, Surrey, UK) (35 µm × 35 µm) by cross-chopping twice. The tissue was then suspended in KRB. Brain minces were incubated at 10°C for 30 min. Minces were washed once and resuspended in 5 ml of KRB containing [³H]adenine (21.0 Ci/mmol; Amersham Biosciences Inc., Piscataway, NJ) (4 µCi/ml) and incubated for 60 min at 32°C under a 95% O₂/5% CO₂ atmosphere. Tissue was then washed twice with fresh KRB and distributed (250 µl) into each well of a 24-well cell culture plate containing KRB buffer, 1 mM 3-isobutyl-1-methylxanthine (a phosphodiesterase inhibitor) (Calbiochem, La Jolla, CA), and stimulant (forskolin or isoproterenol). The assay was carried out for 15 min in a 32°C water bath. The reaction was stopped with 50 µl of 100% (w/v) trichloroacetic acid. Unlabeled cAMP and [-³²P]ATP (PerkinElmer Life Sciences, Boston, MA) were added for recovery calculations (Johnson et al., 1994). After sequential chromatography on Dowex and alumina columns using the method of Salomon et al. (1974), samples were quantitated by liquid scintillation counting. cAMP production was calculated as the percentage conversion of [³H]ATP to [³H]cAMP.

Western Blot Analysis of cAMP Response Element Binding Protein (CREB). The cerebellum was dissected as described above from C57 or DBA mice and was homogenized by sonication in 10 volumes of 2% SDS (two times at 50% for 15 s), and boiled for 5 min. After protein determination (bicinchoninic acid method; Pierce, Rockford, IL), 5-µg aliquots of protein from individual animals were subjected to SDS-polyacrylamide gel electrophoresis using an 8% resolving gel (Novex; Invitrogen, Carlsbad, CA). Proteins were transferred electrophoretically to nitrocellulose membrane (22 µm; Osmonics, Inc., Westborough, MA). Blots were blocked and then incubated with primary antibody (Creb, 1:1000; Cell Signaling Technology, Beverly, MA) as per the manufacturer's instructions with the following exceptions. Blots were washed in Tris-buffered saline with 0.05% Tween. The blots were then incubated with horseradish peroxidase conjugated to goat anti-rabbit IgG for 1 h at room temperature, and protein was visualized using the enhanced chemiluminescence method according to the manufacturer's instructions (Renaissance; PerkinElmer Life Sciences). After exposure to Kodak X-Omat X-ray film (Eastman Kodak, Rochester, NY), blots were stained for total protein using India Ink stain.

Preparation of Brain cDNA and Amplification for Sequencing. Preparation of PCR products from Creb1 cDNA was performed as described by Ehringer et al. (2001). Briefly, C57 and DBA mice were sacrificed by cervical dislocation, and cerebellum was isolated and immediately frozen on liquid nitrogen. Total RNA was obtained from individual animals with the RNAgents Total RNA Isolation System (Promega, Madison, WI). Using the SuperScript First-Strand Synthesis System for reverse transcription-PCR (Invitrogen), cDNA was generated. cDNAs from three C57 or three DBA mice were combined. Primers were designed to amplify overlapping segments of the Creb1 cDNA, each approximately 400 bases. The program Primer3 (Rozen and Skaletsky, 2000, http://www-genome.wi.mit.edu/cgi-bin/primer/primer3www.cgi) was used to identify primer oligonucleotides with the appropriate features: 18 to 22 nucleotides, 40 to 60% GC content, and estimated annealing temperatures of 55-62°C. M13 phage promoter universal tails were added to the 5' ends so that a single primer could be used for sequencing reactions. Polymerase chain reactions were carried out in PerkinElmer GeneAmp PCR System 9600 machines with cycling parameters as follows: 4 min at 94°C, 35 cycles of 30 s at 94°C, 75 s at 58°C, 75 s at 72°C, followed by 10 min at 72°C. Ready-To-Go PCR beads (Amersham Biosciences, Piscataway, NJ) were dispensed into 200-µl tubes, and cDNA, water and primers were added to a final volume of 35 µl. The products were viewed on ethidium-stained 1.8% agarose gels by using UV light. Bands of the appropriate size were cut out of the gel and placed in 1.5-ml tubes, and the PCR products were purified using the QIAquick Gel Extraction Kit (QIAGEN, Valencia, CA).

Sequencing Analysis. PCR products were sequenced directly as described by Ehringer et al. (2001), using the ABI Prism Big Dye Terminator Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA). Sequencing reactions were carried out as follows: 35 cycles of 10 s at 96°C, 50 s at 50°C, and 4 min at 60°C. Precipitated samples were loaded onto an ABI 373 Automated Sequencer or an ABI 3100 Genetic Analyzer (Applied Biosystems). Sequence data were transferred to a UNIX computer, where they were analyzed by using the programs CONSED, Phred, and Phrap (Ewing and Green, 1998; Gordon et al., 1998). A minimum coverage of four times (two forward and two reverse reads) was completed for each section of Creb1 cDNA for each mouse strain. PolyPhred was used to assist in the identification of potential DNA differences (Nickerson et al., 1997). Sequences were submitted to GenBank (http://www.ncbi.nlm.nih.gov/; see Results).

Statistical Analyses. Final analyses for strain means and QTLs are reported for 30 strains of BXD RI mice. Statistical analysis of initial sensitivity (BEC₀), AFT, and cAMP production data was performed using ANOVA. Post hoc analysis used the Tukey test. Genetic correlations were calculated among strain means using Pearson product moment correlation analysis (where r is the correlation). All analyses were performed with SPSS 9.0.1 or SigmaStat 2.03 (SPSS Science, Chicago, IL). p values <0.05 were taken to indicate significant differences in mean values. Estimates of heritability (h²) were calculated from the ANOVA values as h² = SS_B/SS_T, where SS_B equals the sum of squares between subjects and SS_T is the sum of squares total (Belknap, 1998).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Initial sensitivity and AFT were measured in each of the 30 BXD RI strains available and the parental strains (C57 and DBA mice). Both initial sensitivity (BEC₀) and AFT were continuously and quantitatively distributed across strains. Initial sensitivity differed significantly across strains (Fig. 1A, ANOVA p < 0.001). AFT also differed significantly across strains (Fig. 1B, ANOVA p < 0.001). Genotype accounted for 36% and 39% of the total variance observed for initial sensitivity and AFT, respectively (i.e., h² = 0.36, 0.39). Times for fall and regain of balance revealed significant differences between strains (Table 1, ANOVA p < 0.01). Mean (±S.E.M.) time for fall from the dowel across strains was 1.88 ± 0.007 min, and mean times for first regain and second regain of balance on the dowel across strains were 18.96 ± 1.02 and 150.43 ± 2.34 min, respectively. There was a significant correlation between time for loss of balance on the dowel and blood ethanol concentration at fall from the dowel (initial sensitivity; Table 2, r = 0.64, p < 0.01). Blood ethanol concentrations at regain of balance (BEC₁ and BEC₂) correlated significantly with times for regain of balance (regain 1 and regain 2; see Table 2). There was no significant correlation between initial sensitivity (BEC₀) and AFT (Table 2).

View larger version (74K): [in this window] [in a new window]   Fig. 1.   A, blood ethanol concentrations (BEC0) as a measure of initial sensitivity on the dowel test for ataxia in BXD RI, C57, and DBA mice. Bars represent strain means (for n, see Table 1); error bars indicate S.E.M. ANOVA revealed significant differences between strains (p < 0.001). B, acute functional tolerance (AFT) on the dowel test for ataxia in the same BXD RI, C57, and DBA mice as in panel A. AFT is defined as the difference between BEC1 and BEC2 (BEC2  BEC1). Bars represent strain means; error bars indicate S.E.M. ANOVA revealed significant differences between strains (p < 0.001).

Prior to biochemical measures in the BXD RI mice, pilot studies were performed to determine whether behavioral testing would influence or alter biochemical measures in the same animals. Naive C57 and DBA mice were compared with littermates tested on the dowel test for ataxia as described under Materials and Methods. At 1 or 2 weeks after behavioral testing, whole-cell cAMP accumulation was measured in the cerebellum of the behaviorally tested mice as described under Materials and Methods, and the biochemical results were compared with those of naive controls. There was no significant difference in basal, forskolin-, or isoproterenol-stimulated cAMP accumulation in the cerebellum between naive controls and mice tested behaviorally 1 or 2 weeks before determination of cerebellar biochemistry in either the C57 or DBA strain of mice (data not shown).

Basal, forskolin (a direct activator of adenylyl cyclase)-, and isoproterenol-stimulated cAMP accumulation in the cerebellum was measured in the 30 BXD RI strains of mice and the C57 and DBA mice (Fig. 2, A-C). Basal and stimulated cAMP production was continuously and quantitatively distributed across strains. ANOVA revealed significant differences in basal (p < 0.002, Fig. 2A), forskolin (p < 0.001, Fig. 2B)-, and isoproterenol-stimulated (p < 0.001, Fig. 2C) cAMP accumulation in the cerebellum. Genotype accounted for 24, 35, and 42% of the total observed variance for basal, forskolin-, and isoproterenol-stimulated cAMP accumulation, respectively (h² = 0.24, 0.35, 0.42).

View larger version (57K): [in this window] [in a new window]   Fig. 2.   cAMP accumulation measured in cerebellar preparations from the BXD RI, C57, and DBA mice 1 week after behavioral testing shown in Fig. 1, A and B under basal conditions (panel A), in the presence of forskolin (10 µM) (panel B), or stimulated by isoproterenol (1 µM) (panel C). Bars represent strain means for cAMP accumulation (for n, see Table 1); error bars indicate S.E.M. ANOVA revealed significant differences between strains under basal (p < 0.002), forskolin-stimulated (p < 0.001), and isoproterenol-stimulated (p < 0.001) conditions.

We identified a significant genetic correlation between initial sensitivity to ethanol on the dowel test for ataxia and basal cAMP accumulation in the cerebellum (r = 0.433, p = 0.02, Table 3). This negative correlation means that mice with high basal cAMP accumulation (DBA allele) have an increased sensitivity to ethanol, or low BEC₀ (DBA allele). There were no significant correlations between times for loss or regain of balance and basal or stimulated cAMP accumulation measures (Table 3).

Provisional QTLs for initial sensitivity and AFT on the dowel test for ataxia and basal and stimulated cAMP production are presented in Table 4. Multiple provisional QTLs were identified for both behavioral and biochemical measures. Since we identified a genetic correlation between initial sensitivity to ethanol on the dowel test for ataxia and basal cAMP accumulation in the cerebellum, we would expect to find common QTLs for these two phenotypes. Since we found no genetic correlation between initial sensitivity and AFT, we would expect to find few, if any, QTLs in common between these behavioral measures. Overlapping QTLs were identified for initial sensitivity (BEC₀) and basal cAMP accumulation on chromosomes 1 and 8 (Table 4), which suggests that they may share certain genetic codetermination. Overlapping QTLs for basal and forskolin-stimulated cAMP accumulation were also identified on chromosome 17 (Table 4), suggesting that a common gene(s) may influence these biochemical phenotypes.

As mentioned, we were particularly interested in identifying and characterizing genes that were contained in overlapping QTLs for initial sensitivity on the dowel test for ataxia and basal cAMP accumulation in the cerebellum. We chose to examine Creb1 (Chr 1:31.0 cM) as a good candidate gene because it is a downstream effector of the cAMP signaling pathway. Total CREB levels were compared using Western blot analysis between C57 and DBA mice, and no significant differences were detected (Fig. 3, A and B). The coding region of the Creb1 gene was sequenced to identify potential polymorphisms between the two mouse strains. No differences were detected between C57 (AF448507) and DBA (AF448508) nucleotide sequences for the Creb1 cDNA (data not shown). Additionally, we compared the nucleotide sequences we obtained for the Creb1 cDNA to those available in on-line databases from Celera Genomics (11/2001, http://www.celera.com) and the Mouse Genome Sequencing Consortium (11/2001, http://www.ncbi.nlm.nih.gov/genome/seq/MmHome.html) for the DBA and C57 mouse strains, respectively. The Celera Genomics database contains 91% of the coding region of the DBA Creb1 gene. The Mouse Genome Sequencing Consortium database contains 88% of the coding region for the Creb1 gene. Upon comparison, we detected no differences between our results for either strain and those reported by the two databases.

View larger version (33K): [in this window] [in a new window]   Fig. 3.   Western blot analysis of total CREB protein levels in C57 and DBA mice. A, representative blot showing cerebellar CREB immunoreactivity from individual, naive C57 (C) and DBA (D) mice and control (CREB control extracts; Cell Signaling Technology). Blots were stripped and re-probed with anti--actin as a loading control. Standards were included to determine linearity of signal and for comparison across blots. B, summary of Western blot analysis of CREB levels. Blots were standardized using the standard curve and data are reported as arbitrary densitometric units. Values are reported as the mean ± S.E.M. (n = 10 mice per strain). The t test revealed no significant differences between strains in CREB or -actin (actin) protein levels.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Previously, we identified a genetic correlation between initial ataxic sensitivity to ethanol using the stationary dowel and cAMP signaling in the mouse cerebellum using nine inbred strains of mice (Kirstein and Tabakoff, 2001). Phenotypic differences between inbred strains of mice, assessed under similar environmental conditions, can be interpreted as evidence of genetic influence on the phenotype (McClearn, 1991). A genetic correlation suggests that a common gene, or genes, influences both phenotypic measures. However, genetic correlations fail to identify the actual genes that influence the variation in complex traits. QTL analysis allows for narrowing of the location of genes that may influence the phenotype of interest. Given a genetic correlation between phenotypic traits, one would predict that there would be overlapping QTLs for these traits. One would also expect that the gene(s) influencing both the genetically correlated traits would be contained within the overlap of the QTL. There are several advantages of utilizing the BXD RI strains compared with other RI strains for QTL mapping. First, an elaborate chromosomal marker map exists (http://www.informatics.jax.org; Blake et al., 2001), and additional markers have been typed in the BXD RI strain panel (Taylor et al., 1999), improving marker map density and the ability to detect associations. Likewise, many investigators use the BXD RI panel of mice, making it possible also to compare results across studies and identify genetic correlations (for review see Wehner et al., 2001). Furthermore, additional strains have been added to the BXD RI panel (Taylor et al., 1999), which increases the power to detect associations between complex traits and genetic markers.

We detected significant differences between strains in initial sensitivity, defined as BEC₀, on the dowel test for ataxia (blood ethanol concentrations at the time a mouse was no longer able to balance on a dowel) using the BXD RI panel of mice. There are various behavioral tests of ethanol-induced ataxia. They include the stationary dowel, the Rotarod, the screen test, and the grid test involving measures of grip strength, balance, and coordinated movement (Gallaher et al., 1996; Phillips et al., 1996; Browman and Crabbe, 2000; Deitrich et al., 2000; Gehle and Erwin, 2000). Although inbred strains of mice have been shown to differ on numerous measures of ethanol sensitivity, no one strain has been shown to be more or less sensitive across all tests of ethanol sensitivity (Crabbe, 1983; Crabbe et al., 1994). Boehm et al. (2000) demonstrated that 5-hydroxytryptamine-1B mutants differed from their wild-type littermates on only two of eight different measures of motor incoordination. In other words, all of the "sensitivity" measures are not genetically correlated to one another. We were unable to identify any genetic correlation between our results of initial sensitivity using the stationary dowel and the results of Gallaher et al. (1996) for initial sensitivity using the Rotarod. Crabbe et al. (1996) were also unable to find a genetic correlation between initial sensitivity using the Rotarod and the grid test. The lack of genetic correlation across such studies further points to the specificity of genetic influences on behavioral tests of initial sensitivity to ethanol-induced ataxia.

Gehle and Erwin (2000) evaluated the LSxSS RI panel of mice using the same behavioral test (the dowel test for ataxia) that we have employed in our study to identify QTLs for initial sensitivity and AFT. The QTLs for initial sensitivity identified by Gehle and Erwin (2000) were unlike those identified in our studies. One possible reason we did not find overlapping QTLs between their study and our results may be related to a difference in the composition of the genetic pool from which the BXD RIs and the LSxSS RIs were derived. The LSxSS RI mice were originally derived from a heterogeneous stock of eight inbred strains (McClearn and Kakihana, 1981). The DBA and C57 mice were two of the progenitor strains for the LSxSS RI mice, but the effects of unique alleles contributed by the other six inbred progenitor strains to the LS and SS mice may explain the lack of overlap between QTLs identified from the two studies. The heritabilities, which are attributed to a specific population, for initial sensitivity using the dowel test for ataxia were found to differ in the BXD RI mice and the LSxSS RI mice (our study and Gehle and Erwin, 2000).

Interestingly, in our studies and two other studies using BXD RI mice, overlapping QTLs were identified between initial sensitivity on the dowel test and the screen test for ataxia on chromosome 1 (74.5 and 93.3 cM), chromosome 8 (13 cM), and chromosome 11 (26-32 cM) (Browman and Crabbe, 2000), and with initial sensitivity on the grid test on chromosome 1 (93 cM) and chromosome 11 (35-40 cM) (Phillips et al., 1996). These findings suggest possible genetic codetermination for certain forms of ataxia in the BXD RI mice.

When we examined our QTL results for AFT and those of Gallaher et al. (1996), who also used BXD RI mice, we were unable to identify any genetic correlation between our results for AFT using the stationary dowel and the results of Gallaher et al. (1996) for acute tolerance (delta or fold-increase) using the Rotarod. This is not completely unexpected because, as mentioned above, there is little evidence that genetic correlations exist between measures of ethanol's behavioral actions on a Rotarod and ethanol's behavioral actions measured by use of a stationary dowel or a grid (current study, and Crabbe et al., 1996).

Deitrich et al. (2000) and Gehle and Erwin (2000) used the LSxSS RI mice and the dowel test for ataxia to identify QTLs for AFT. When comparing our results to those of Deitrich et al. (2000) and Gehle and Erwin (2000), we were only able to identify one common QTL between our study and the study by Gehle and Erwin (2000) for AFT. This region occurred on chromosome 14 at 0.5 cM (in our study) and at 3 cM (in the study of Gehle and Erwin, 2000). Again, the lack of overlapping QTLs in studies using BXD RIs versus the LSxSS RIs may be due to differences between the BXD RI and the LSxSS RI mice in their gene pool and in the genetic pathways used to arrive at AFT to ethanol.

Cyclic AMP signaling has been implicated in the ataxic effects of ethanol (see above) and in the development of alcohol tolerance (for review see Tabakoff and Hoffman, 1998). Using BXD RI mice, we identified a significant genetic correlation between initial sensitivity on the dowel test for ataxia and basal cAMP accumulation in the cerebellum. We did not find a genetic correlation between AFT and cAMP accumulation in the cerebellum. Our previous work also supports the lack of a genetic relationship between cAMP accumulation and AFT. High acute functional tolerance mice and low acute functional tolerance mice, selectively bred for differences in AFT using the stationary dowel (Erwin and Deitrich, 1996), did not differ in cAMP accumulation in the cerebellum (Kirstein and Tabakoff, 2001).

In accord with the significant genetic correlation between initial sensitivity using the dowel test and basal cAMP accumulation in the cerebellum, we identified two overlapping QTLs for these measures. In general, strains possessing the DBA allele were more sensitive to ethanol-induced ataxia (i.e., the mice fell from the dowel at a lower BEC) and had a higher basal cAMP accumulation measure. Conversely, strains possessing the C57 allele were less sensitive to ethanol-induced ataxia (i.e., the mice fell from the dowel at a higher BEC) and had a lower basal cAMP accumulation measure. Our results, of course, do not rule out the possibility that brain regions other than the cerebellum may be involved, since the mice are mobile on the dowel, actively moving along its length, and other pathways such as basal ganglia might play a role in the incoordinating actions of ethanol (Kirstein and Tabakoff, 2001). Verification of our provisionally identified QTLs in a genetically segregating population will also be required to conclusively answer which of the QTLs represent true genetic influences. It should be noted, however, that the actual level of significance for many of the QTLs identified in our studies (Table 4) was in the p < 0.001 to p < 0.005 range. Such data provide greater confidence in the association of the provisional QTLs and the measured phenotypes. The provisional QTLs are thus useful in generating hypotheses about candidate genes, as well as in focusing future studies with other genetic models.

Genes were considered candidates within the context of our studies if they mapped within ±10 cM of markers identifying a particular QTL. Possible candidate genes that map near QTLs for initial sensitivity include the cAMP response element binding protein 1 (Creb1, Chr 1:31.0 cM), glycine receptor ₁ (Chr 11:30.0 cM), glutamate receptor ionotropic -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid 1 (Chr 11:31.0 cM), acetylcholine receptor  (Chr 11:40.0 cM), acetylcholine receptor  (Chr 11:42.0 cM), and the 5-hydroxytryptamine (serotonin) transporter (Chr 11:42.0 cM) (www.informatics.jax.org; Blake et al., 2001). Candidate genes of interest that map near-provisional QTLs for basal cAMP accumulation measures include adenylyl cyclase 1 (Chr 11:1.25 cM), cAMP response element binding protein 1 (Chr 1:31.0 cM), adenylyl cyclase 7 (Chr 8:40 cM), and calcium/calmodulin-dependent protein kinase II- (Chr 11:0.5 cM) (www.informatics.jax.org; Blake et al., 2001). Interestingly, the overlapping QTL for both initial sensitivity and basal cAMP accumulation on chromosome 1 contained Creb1 (Chr 1:31.0 cM) (http://www.informatics.jax.org; Cole et al., 1992; Blake et al., 2001). CREB is a major transducer of cAMP signals and through its transcriptional control function (Andrisani, 1999; Shaywitz and Greenberg, 1999) could influence basal gene expression levels which, in turn, could impact an animal's response to ethanol (Wand et al., 2001).

To investigate the possibility that Creb1 may be the gene influencing initial sensitivity and basal cAMP accumulation, we compared CREB levels and gene sequence (cDNA) between the parental strains (C57 and DBA). Differences in CREB expression levels could imply sequence differences in the promoter region of the gene. Identifying sequence differences in the coding region of the gene could implicate allelic differences in CREB function. We, however, found no differences in expression levels or coding sequence of Creb1. The Creb1 mRNA can be alternatively spliced to produce several CREB isoforms (Ruppert et al., 1992; Blendy et al., 1996). Since our methods did not include the sequencing of intronic regions of the Creb1 gene, we cannot rule out the possibility of differential alternative splicing in the C57 and DBA mice.

Many genes lie within the QTLs from which we chose Creb1 as a candidate gene. There are at least 70 known genes in the QTL interval on chromosome 1 reported on the Mouse Genome Informatics web site (http://www.infomatics.jax.org; Blake et al., 2001), but there could be even more genes based on empirical estimates. With advances in genomic information for the mouse, one can search for polymorphisms using on-line databases provided by Celera Genomics (http://www.celera.com) and the Mouse Genome Sequencing Consortium (http://www.ncbi.nlm.nih.gov/genome/seq/MmHome.html) to obtain mouse genomic sequence data for in silico comparative sequencing between strains. Our sequencing data for Creb1 were obtained prior to access to the Celera Genomics database. Current comparison of our sequence results with the database for DBA (http://www.celera.com) and C57 (http://www.ncbi.nlm.nih.gov/genome/seq/MmHome.html) mice indicated that no significant differences exist between our data and the regions containing the coding regions of the Creb1 gene in the Celera Genomics or the Mouse Genome Sequencing Consortium databases.

Our results demonstrate a significant genetic influence on initial sensitivity (BEC₀) and AFT using the dowel test for ataxia and also for basal, forskolin-, and isoproterenol-stimulated cAMP accumulation in the cerebellum. We were able to detect multiple provisional QTLs for both behavioral and biochemical phenotypes. The identification of a significant genetic correlation between initial sensitivity (BEC₀) using the dowel test for ataxia and basal cAMP accumulation in the cerebellum was further supported by the detection of two overlapping QTLs for these measures. The identity of genes that contribute to behavioral and biochemical differences is currently being pursued through comparative sequencing and differential gene expression in our laboratories using on-line genomic sequence databases and gene expression array technology, respectively.