The D5-/- and their D5+/+ and D5+/- littermates were produced as previously described (Holmes et al., 2001). The mice from this mutant line were backcrossed with C57BL/6 mice for 10 generations before testing. We utilized a heterozygous breeding scheme in order to produce mixed litters with all three genotypes. Mouse genotypes were confirmed by PCR. Mice were weaned at P28 and group housed except in the T-maze experiments where mice were single housed starting 1 week before testing. Mice used for testing were male and between P63 and P126 days of age. All procedures were approved by the National Institute of Mental Health Animal Care and Use Committee and followed the National Institutes of Health Using Animals in Intramural Research Handbook.

The procedure for this T-maze task was similar to one previously used in our laboratory (Papaleo et al., 2008). Mice were habituated to single housing for 1 week and were then food restricted to a level of 85% of their free-feeding weight. The mice were given 8 days for their weight to stabilize and received access to 10 reward pellets (5TUL 14 mg pellets; TestDiet, Richmond, IN, United States) during the last 3 days of this period. Following habituation to single housing and stabilization of body weight, mice were habituated to the T-maze apparatus over the course of two sessions. The T-maze apparatus was made of clear acrylic [dimensions of arms (length × width × height): 40 × 10.2 × 17.5 cm]. A recessed food cup was located at the end of each arm. During habituation sessions mice were allowed to retrieve reward pellets from the food cups. At the beginning of the session, each cup was baited with two reward pellets. The cups were re-baited continuously. Mice were allowed to retrieve 16 reward pellets during Session 1 and 20 reward pellets during Session 2. Each session automatically ended after 10 min if the mouse did not retrieve the maximum number of reward pellets. On the day following habituation, mice were given one session of 10 forced-alternation runs. For this session, one goal arm was blocked and the mouse had 2 min to consume the reward pellet located in the open arm. After an inter-trial interval of at least 15 min, the mouse was returned to the maze for another forced run with the open/closed arms switched. Training for the discrete paired-trial delayed alternation task began on the following day. Training consisted of 10 paired trials each day. A paired trial consisted of a forced run where one arm was blocked and the other arm was baited with a single reward pellet. The mouse was given 4 min to consume the pellet. Following consumption, the mouse was returned to the home cage for a 4-s intratrial delay. After the intratrial delay, the mouse was returned to the maze with access to both arms. The arm blocked on the forced run was now baited with two reward pellets. Again, the mouse was given 4 min to consume the reward pellets. After an inter-trial interval of at least 15 min, mice were returned for another trial. If the mouse entered the unbaited arm, this was recorded as an error and the mouse was removed from the maze. The normal inter-trial interval followed incorrect trials as well. Each testing session utilized a pseudo-randomly chosen pattern of 10 forced runs. Each day, the same pattern was used for each mouse. Mice were trained using these parameters for 20 days or until they reached 80% accuracy for 3 consecutive days. Mice that failed to reach 80% accuracy for 3 consecutive days within the 20-day training period were excluded from the study. Mice were then tested using variable intratrial intervals (4, 30, 60, and 240 s) and a 20-s inter-trial interval. Mice were given four trials of each inter-trial interval on 4 consecutive days.

The experimental apparatus consisted of a novel Plexiglas open field arena (42 × 42 × 30 cm) under red light illumination (5 ± 2 lux). Each mouse was allowed to freely explore the open field alone for 60 min. Horizontal locomotor activity was recorded using infrared photobeam sensors and the VersaMax Open Field Activity Monitoring system (AccuScan Instruments, Inc., Columbus, OH, United States).

Temporal order object recognition testing was conducted as previously described (Managò et al., 2016). The apparatus and lighting conditions were identical to those used for the open field locomotor activity test. On day 1, mice were allowed to freely explore the open field for 60 min. Day 2 consisted of three 5-min sessions. During the first session, mice were allowed to explore two identical objects within the open field arena. The objects were either rectangular boxes (3 × 3 × 6 cm) or Erlenmeyer flasks (4 × 6 cm). The objects could either be white or black. During the second session, 1 h after the first session, mice were allowed to explore two objects of a different shape and color with respect to the objects from the first session. During the third session, 3 h after the second session, mice were allowed to explore one copy of each of the objects presented during the first and second sessions. The sessions were videotaped and scored offline by a reviewer blind to genotype. Mice were considered to be exploring an object when they faced the object and were ≤2 cm from the object. Discrimination between the objects in the third session was calculated using a discrimination index that accounts for individual differences in total exploration time. The index was calculated as the difference between the time spent exploring the object from the first session and the object from the second session divided by the total exploration time. Any mice that did not explore objects for more than 4 s during all of the sessions were excluded from the final analysis. One D5+/+, one D5+/-, and four D5-/- mice were excluded because of low total exploration.

Frontal cortex tissue was obtained from naive D5+/+, D5+/-, and D5-/- mice. Briefly, mice were killed by decapitation and the brain was removed and placed on a glass tray on ice. The olfactory bulbs were removed and the brain was cut along the midline. The frontal cortex (anterior to the corpus callosum) was removed and flash frozen on dry ice. The tissue was then stored at -80°C until processing. The tissue was then homogenized and sonicated in T-Per (Thermo Scientific, Rockford, IL, United States) lysis buffer. The protein concentration of the samples was determined by Bradford assay and all samples were diluted to a final concentration of 2 μg/μl. The samples were then combined with NuPage^® LDS Sample Buffer (ratio of 3:1 sample:LDS buffer). The protein was then denatured by heating at 95°C for 5 min. Samples were then run on pre-cast 4–12% Bis-Tris gels. After transfer to PVDF membranes, blots were blocked with 5% milk in TTBS for 1 h at room temperature. We probed the blots with the primary antibodies at 4°C overnight (Table 1). Following three 10-min washes in TTBS, blots were probed with the goat anti-rabbit, goat anti-mouse (1:10,000 dilution; Chemicon, Temecula, CA, United States), or donkey anti-goat (1:1000; Santa Cruz Biotechnology, Dallas, TX, United States) secondary antibodies at room temperature for 1 h. Blots were developed in ECL-Plus (GE Healthcare, Piscataway, NJ, United States) and exposed to Kodak Bio-Max film. Films were digitized using a scanner, and the resulting images were analyzed using NIH Image gel plotting macros.

The habituation and training phases of the T-maze task were analyzed using one-way ANOVAs with post hoc Bonferroni’s tests. The variable delay portion was analyzed by a two-way ANOVA with genotype as the between-subjects factor and retention interval as the within-subjects factor and post hoc analyses utilized Bonferroni’s tests at each of the retention intervals. Data from the temporal order object recognition task were analyzed using a one-way ANOVA with post hoc Bonferroni’s tests. Protein quantification data for each mouse was first normalized to GAPDH levels (actin for TH protein levels) and then normalized to the mean D5+/+ value for each protein. Group means were then compared using one-way ANOVAs with post hoc Bonferroni’s tests. All data are shown as the mean ± SEM.

To identify any potential confounding effects of D₅ receptor disruption, we measured open field locomotor activity and approach responses to food reward in the T-maze apparatus. There were no D₅ genotype differences in total locomotor activity (F_(2,30) = 0.06, p = 0.9420; Figure 1A) or any genotype × time interactions during any of the 5-min time bins within the 60-min test (F_(22,330) = 0.63, p = 0.9000; Figure 1B). As expected there was a significant decrease in activity over time for all groups (F_(11,330) = 35.68, p < 0.0001; Figure 1B).

The first phase (two sessions) of the T-maze task is designed to habituate the mice to the testing apparatus and retrieval response required for completion of the task. All groups showed a significant decrease in latency to consume the first pellet from Day 1 to Day 2 (F_(1,30) = 41.71, p < 0.0001; Figure 2A). Additionally, there were no differences in the raw latency values between the genotypes on either Day 1 or Day 2 (F_(2,30) = 1.20, p = 0.31). These data support the notion of a negligible impact of D₅ on measures of locomotion and motivation.

The next stage of the T-maze task consisted of training required for the mice to learn the non-match to sample rule. There were no significant differences in the number of sessions required to reach the criterion of three consecutive sessions above 80% accuracy (F_(2,29) = 0.70, p = 0.50; Figure 2B). However, there was a significant difference in accuracy during those last three training sessions (F_(2,29) = 3.63, p = 0.04; Figure 2C). The genotype effect on accuracy was driven by a slight, but significant, difference between D5+/- and D5-/- (p = 0.04). The difference between D5+/+ and D5-/- was not significant (p = 0.22). These data indicate a marginal role of the D₅ receptor in the ability to acquire working memory rules.

Following successful completion of the training phase, mice were tested in the variable retention delay portion of the task. In addition to the variable retention delays, the inter-trial delay was set at 20 s (Aultman and Moghaddam, 2001). Accuracy decreased as the retention interval increased across all genotypes (F_(2,29) = 11.00, p < 0.0001). There was also a significant main effect of D₅ genotype on performance (F_(2,29) = 4.91, p = 0.01). D5-/- mice showed significantly impaired performance compared to D5+/+ on trials with either 4- or 30-s retention intervals (p = 0.004 and 0.001, respectively; Figure 2D). D5+/- mice had impaired performance at the 30-s interval compared to D5+/+ mice (p = 0.033; Figure 2D). There were no significant differences between the genotypes on either the 60- or 240-s retention intervals due to a decrease in the choice accuracy of the D5+/+ mice. Interestingly, the D5+/- mice had an intermediate phenotype suggesting that there is a gene dose effect of D₅ dopamine receptor expression. These findings highlight a clear and previously undetected role of D₅ receptors in working memory abilities.

Performance in a temporal order object recognition task, like the non-match to sample T-maze task, has been shown to depend on intact PFC function (Barker et al., 2007). There were no differences in total exploration between the three genotypes (F_(2,31) = 0.12, p = 0.89; Figure 3A), but there was a significant difference in the discrimination index (F_(2,31) = 3.33, p = 0.049; Figure 3B). Post hoc analysis indicated that the D5+/- group had a significantly lower discrimination index compared to the D5+/+ group (p = 0.046), but only a tendency was evident for D5-/- mice. Thus, there might be an U-shaped gene-dose effect on temporal order object recognition as the D5-/- group’s discrimination index was not different from the D5+/+ group’s index (p = 0.52). One sample t-tests indicated that both the D5+/+ (t₍₇₎ = 3.28, p = 0.01) and D5-/- (t₍₁₀₎ = 2.50, p = 0.03) groups showed significant recency memory while the D5+/- (t₍₁₄₎ = 0.44, p = 0.67) group did not.

The working memory and temporal order recognition deficits exhibited by D₅ mutant mice are similar to those our group has observed in other mouse genetic models of dopamine-related cognitive dysfunction (Papaleo et al., 2008, 2012). Moreover, performance in the discrete paired-trial variable-delay T-maze and temporal order object recognition tasks has been shown to be modulated by alterations in PFC function (Kellendonk et al., 2006; Barker et al., 2007). Thus, we next investigated whether D₅ knockout mice might have working memory- and dopamine-related molecular alterations within the mPFC (Papaleo et al., 2008, 2014; Tan et al., 2012; Easton et al., 2013; Managò et al., 2016).

We measured the relative amounts of multiple proteins in the frontal cortex across all three genotypes. These data are presented in Figure 4. Akt is a key intracellular regulatory protein involved in dopaminergic signaling and implicated in psychiatric disorders (Emamian et al., 2004; Beaulieu et al., 2007). There was no significant difference in pAkt Thr308 (F_(2,22) = 0.46, p = 0.63; Figures 4A,B) or total Akt (F_(2,22) = 0.46, p = 0.63). However, there was a significant genotype effect on pAkt Ser473 protein levels (F_(2,22) = 4.90, p = 0.02) with D5-/- mice exhibiting increased phosphorylation compared to D5+/+ mice (p = 0.02). Despite seeing large changes in CAMKKβ protein levels in other genetic mouse models with working memory deficits (Papaleo et al., 2008, 2012), there were no changes in the CAMKKβ 66 kDa (F_(2,22) = 0.03, p = 0.97) and a small, but not statistically significant, decrease in D5-/- mice compared to D5+/+ on the 64 kDa (F_(2,22) = 2.315, p = 0.12) isoform. Protein levels of another Ca²⁺-dependent kinase CAMKII were not significantly different between the genotypes (F_(2,22) = 0.07, p = 0.94). Also, protein levels of the CAMKKβ substrate CAMKIV were unaffected by the D₅ genotype (F_(2,22) = 0.31, p = 0.73). Finally, we measured levels of dopamine-associated proteins and found no significant differences in the D₂ receptor (F_(2,22) = 1.19, p = 0.32), membrane-bound catechol-O-methyltransferase (COMT) (F_(2,22) = 0.65, p = 0.53), soluble COMT (F_(2,22) = 1.79, p = 0.19), phosphorylated (Ser40) tyrosine hydroxylase (TH) (F_(2,13) = 0.01, p < 0.99), or TH total protein levels (F_(2,13) = 1.04, p < 0.38; Figures 4A,B). Overall, these findings indicate a selective impact on PFC Akt activation by D₅ receptor while sparing other dopamine-related biomarkers.