Region-specific deletions of RIM1 reproduce a subset of global RIM1α−/− phenotypes

The presynaptic protein RIM1α mediates multiple forms of presynaptic plasticity at both excitatory and inhibitory synapses. Previous studies of mice lacking RIM1α (RIM1α−/− throughout the brain showed that deletion of RIM1α results in multiple behavioral abnormalities. In an effort to begin to delineate the brain regions in which RIM1 deletion mediates these abnormal behaviors, we used conditional (floxed) RIM1 knockout mice (fRIM1). By crossing these fRIM1 mice to previously characterized transgenic cre lines, we aimed to delete RIM1 selectively in the dentate gyrus (DG), using a specific preproopiomelanocortin promoter driving cre recombinase (POMC-cre) line , and in pyramidal neurons of the CA3 region of hippocampus, using the kainate receptor subunit 1 promoter driving cre recombinase (KA-cre). Neither of these cre driver lines was uniquely selective to the targeted regions. In spite of this, we were able to reproduce a subset of the global RIM1α−/− behavioral abnormalities, thereby narrowing the brain regions in which loss of RIM1 is sufficient to produce these behavioral differences. Most interestingly, hypersensitivity to the pyschotomimetic MK-801 was shown in mice lacking RIM1 selectively in the DG, arcuate nucleus of the hypothalamus and select cerebellar neurons, implicating novel brain regions and neuronal subtypes in this behavior.

In mice, global deletion of presynaptic proteins often results in significant behavioral abnormalities reminiscent of human cognitive disease, though an understanding of brain regions or cell types mediating such phenotypes are not clear. For example, post-mortem studies on schizophrenic brains have consistently found alterations in presynaptic proteins, particularly genes involved in GABAergic neurotransmission, in the cerebellum, hippocampus and cortex (Akbarian and Huang 2006;Eastwood et al. 2001;Glantz & Lewis 1997;Guidotti et al. 2000;Sawada et al. 2005;Tamminga et al. 2004;Woo et al. 2004). Neurexin, another presynaptic protein, has been associated with autism (Etherton et al. 2009;Feng et al. 2006;Kim et al. 2008).
Recently, we have found that deletion of the presynaptic protein Rab3a interacting molecule 1α (RIM1α) in mice induced significant behavioral abnormalities, including some schizophrenia-relevant phenotypes, though RIM1α has not been directly implicated in human schizophrenia. These abnormalities include learning deficits (Powell et al. 2004), decreased prepulse inhibition (Blundell et al. 2010b), increased locomotor response to novelty (Powell et al. 2004), deficits in social interaction (Blundell et al. 2010b), increased sensitivity to the non-competitive N-methyl-Daspartate receptor (NMDAR) antagonist MK-801 (Blundell et al. 2010b) and deficits in maternal behavior . RIM1α was previously shown to express in most neurons (Schoch et al. 2006) and has been shown to mediate specific forms of presynaptic plasticity in hippocampal mossy fibers [dentate gyrus (DG) to CA3 synapses], Schaffer collaterals (CA3 to CA1 synapses), cerebellar parallel fibers (granule cell to purkinje synapses) and GABAergic interneurons (Calakos et al. 2004;Castillo et al. 2002;Chevaleyre et al. 2007;Schoch et al. 2002;Yang & Calakos 2010).
Given RIM1's diverse functions and expression, attributing behavioral phenotypes in the RIM1α knockout mouse to the loss of RIM1α from specific neurons is difficult. Consequently, we generated a conditional (floxed) RIM1 knockout (fRIM1) and crossed it to previously characterized transgenic cre recombinase lines. Initially, we aimed to examine the behavioral effects of deleting RIM1 selectively in DG granule neurons, where it is required for presynaptic long-term potentiation in mossy fibers . Additionally, we aimed to examine the effect of RIM1 deletion selectively in area CA3 pyramidal neurons, where it mediates multiple forms of presynaptic plasticity (Calakos et al. 2004;Schoch et al. 2002). Unfortunately, neither cre driver line was as selective when crossed to our fRIM1 mutants as had been previously reported (Balthasar et al. 2004;McHugh et al. 2007;Nakazawa et al. 2002). Indeed, regional selectivity results of a reporter gene did not correlate completely with our regional measurements of RIM1 mRNA levels.
Nevertheless, we observed that a subset of behaviors were altered in these two regionally restricted RIM1 conditional deletion lines, thereby narrowing the brain regions involved in some behaviors. Specifically, mice lacking RIM1 selectively in the DG, arcuate nucleus of the hypothalamus and select neurons of the cerebellum resulted in increased sensitivity to the psychotomimetic drug MK-801. Our findings also suggest that loss of RIM1 in other brain regions or in multiple brain regions simultaneously may partially reproduce other behavioral abnormalities observed in RIM1α −/− mice.

Genetic manipulations
Floxed RIM1 (fRIM1) mice for genetic removal of RIM1α and RIM1β were generated previously (Kaeser et al. 2008). Briefly, upon homologous recombination of the RIM1αβ targeting vector, 129Sv R1 stem cells containing the fRIM1 construct were injected into C57BL/6J blastocysts to produce chimeric mice, which were crossed for one generation to C57BL/6J for germline transmission. They were then crossed to flp recombinase mice (which were generated by injecting the flp transgene-containing vector into a fertilized egg from a B6SJLF 1 /J X B6SJLF 1 /J cross ;Dymecki 1996) and the resultant offspring were intercrossed to generate the homozygous fRIM1 mice. The wild-type (WT), floxed, and recombined RIM1 alleles were genotyped by polymerase chain reaction (PCR) with oligonucleotide primers as described previously (Kaeser et al. 2008). The preproopiomelanocortin promoter driving cre recombinase (POMC-cre) mouse was generously provided by Joel Elmquist; it was generated in FVB/NJ mice and previously backcrossed three times to C57BL/6J as previously reported (Balthasar et al. 2004;McHugh et al. 2007). The kainate receptor subunit 1 promoter driving cre recombinase (KA-cre) mouse was generously provided by Susumu Tonegawa; it was generated in C57BL/6J mice, crossed to the Roas26 reporter mouse (Soriano 1999) and then subsequently backcrossed eight times to C57BL/6J mice as previously reported (Nakazawa et al. 2002). To generate the fRIM1/POMC-cre and fRIM1/KA-cre mice with sex-matched littermate controls, we used the following three-step breeding strategy. (1) The POMC-cre or KA-cre mice were crossed with fRIM1 homozygous mice. (2) The resulting fRIM1 heterozygous, cre positive mouse from cross 1 was crossed again with fRIM1 homozygotes. (3) The resulting fRIM1 homozygous, cre positive mouse from cross 2 was crossed again with fRIM1 homozygotes. This final cross provided sex-matched littermate pairs on a mixed genetic background that were fRIM1 homozygous, cre positive (fRIM1/cre+, experimental) and fRIM1 homozygous, cre negative (fRIM1/cre-, littermate control). Mice that displayed cre-mediated recombination in the tail DNA were excluded from our study.

Quantitative reverse transcriptase polymerase chain reaction
Rapid tissue harvesting for quantitative reverse transcriptase PCR (qRT-PCR) was performed in chilled dissection solution [26 mM NaHCO 3 , 212.7 mM sucrose, 2.6 mM KCL, 1.23 mM NaH 2 PO 4 , 10 mM dextrose, 3 mM MgCl 2 •6(H 2 O), 1 mM CaCl 2 •2(H 2 O)] and stored at −80 • C until mRNA was extracted with Trizol (Invitrogen). Reverse transcription of the mRNA library was performed using Superscript III 1st strand synthesis kit (Invitrogen). SYBR GreenER qPCR SuperMix Universal Kit (Invitrogen) was used for qRT-PCR according to the manufacturer's instructions. Each tissue sample was tested in triplicate and their Ct values were averaged together. Oligonucleotide primers were designed to create a 105-bp product spanning portions of exons 5 and 6 of the RIM1 mRNA transcript (forward primer -AAGCAGGCATCAAGGTCAAG, reverse primer -ACGTTTGCGCTCACTCTTCT). The neuron-specific microtubule-associated protein 2 (MAP2) was used as a reference gene (Mukaetova-Ladinska et al. 2002;Sundberg et al. 2006) (forward primer -ACTTGACAATGCTCACCACGTA, reverse primer -CCTTTGCATGCTCTCTGAAGTT). The relative change in RIM1 transcript between experimental and control tissue was normalized to MAP2 transcript levels using standard calculation methods (Pfaffl 2001). To test if the normalized relative change in RIM1 transcript levels was significantly different from unity (no change), a one-sample t-test was employed.

Behavioral overview
Mice were housed two per cage (always littermate pairs) and maintained on a light/dark cycle with light on from 0600 to 1800 hours in a 22 ± 2 • C (30-70% humidity) housing room. All behavioral tests were performed in the afternoon. For both fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice, fRIM1/cre− sex-matched littermates were used as WT controls to ensure a constant genetic background. Food (normal chow) and water were available ad libitum for the duration of the behavioral battery. Sex-matched littermate pairs ranged from 3 to 8 months of age at the onset of behavioral testing and all behaviors were completed within 8 weeks. Two cohorts of fRIM1/POMC-cre (N = 12 and N = 8 littermate pairs, total N = 20 littermate pairs; littermate pairs age range at onset of behavior: 3 months, 3 pairs; 4 months, 3 pairs; 6 months, 10 pairs; 7 months, 1 pair; 8 moths, 3 pairs) and two cohorts of fRIM1/KAcre (N = 10 and N = 10 littermate pairs, total N = 20 littermate pairs; littermate pairs age range at onset of behavior: 3 months, 5 pairs; 4 months, 4 pairs; 5 months, 4 pairs; 6 months, 3 pairs; 7 months, 4 pairs) were tested separately and pooled for analysis (except for rotarod, startle threshold, PPI and locomotor response to psychotomimetics in which only one of the cohorts was tested). In total, 13 male and 7 female fRIM1/KA-cre littermate pairs and 9 male and 11 female fRIM1/POMC-cre littermate pairs were tested. Behaviors were performed in the following order: elevated plus maze, dark/light, open field, locomotor, social interaction in the open field, accelerating rotarod, startle threshold, prepulse inhibition, fear conditioning, Morris water maze (MWM) and locomotor response to MK-801. Mice were allowed 1 h to habituate to the testing room before beginning experiments. All tests were carried out in accordance with Institutional Animal Care and Use Committee (IACUC) and UT Southwestern Medical Center animal guidelines and protocols.
Elevated plus maze Elevated plus maze was performed essentially as described previously (Powell et al. 2004). Mice were placed in the center of a black plexiglass elevated plus maze (each arm 33 cm in length and 5 cm wide, with 25-cm-high walls on closed arms) in a dimly lit room for 5 min. Two mazes were used and video-tracked simultaneously (Ethovision 2.3.19, Noldus, Wageningen, The Netherlands). A barrier was set between the mazes to prevent mice from seeing each other.
Time spent in open and closed arms, number of open and closed entries, and time in the middle was calculated. Five fRIM1/POMC-cre and one fRIM1/KA-cre pairs were excluded from analysis because either the experimental or control mouse fell from the platform during testing.
Dark/light Performed essentially as described previously (Powell et al. 2004). Apparatus is a two-compartment opaque plexiglass box (25 × 26 cm in each compartment). One side is black and kept closed and dark, while the other is white with a fluorescent light directly above its open top (1700 lx). Mice were placed in the dark side for 2 min, then the divider between the two sides was removed allowing the mouse to freely explore both chambers for 10 min. Anxiety-like behavior is measured as latency to enter the light side, as well as time spent in the light versus dark compartments. Measures were taken using photobeams and MedPC software (Med Associates, St Albans, VT, USA).

Open field
Performed essentially as described previously (Powell et al. 2004). The open field test was performed for 10 min in a brightly lit, 48 × 48 × 48 cm white plastic arena using the Ethovision video-tracking software (Noldus). Time spent in the center zone (15 × 15 cm) and frequency to enter the center was recorded. Locomotor activity was also measured during the open field test.

Locomotor
Mice were placed in a fresh home cage with minimal bedding for a 2h testing period. Lengthwise horizontal locomotor activity was tested in 5-min bins for the duration of the task using photobeams linked to computer data acquisition software (San Diego Instruments, San Diego, CA, USA). Beams were organized linearly along one horizontal axis in 5-cm increments and total beam breaks were used as the dependent variable.

Social interaction in the open field
Performed essentially as described previously (Blundell et al. 2010a). The test was performed in a 48 × 48 cm white plastic arena under red light using a 6.0 × 9.5 cm perforated plexiglass rectangular cage containing an unfamiliar adult mouse, allowing olfactory, visual, and minimal tactile interaction. Mice were first placed in the arena for 5 min with the empty clear plexiglass cage. Then mice were allowed to interact with a novel caged social target for another 5 min. Time spent in the interaction zone was obtained using Ethovision video-tracking software (Noldus).
Accelerating rotarod Performed essentially as described previously (Powell et al. 2004). A five-lane accelerating rotarod designed for mice (IITC Life Science, Woodland Hills, CA, USA) was used (rod diameter was 3.2 cm; rod length was 10.5 cm). The rotarod was activated after placing mice on the motionless rod. The rod accelerated from 0 to 45 revolutions per min in 60 seconds. Time to fall off the rod was measured.

Startle threshold and prepulse inhibition
Both these tasks were performed as described previously (Etherton et al. 2009). The prepulse inhibition (PPI) test began with six presentations of a 120-dB pulse to calculate the initial startle response. Afterward, the 120-dB pulse alone, prepulse/pulse pairings of 4, 8 or 16 dB above 70-dB background followed by the 120-dB pulse with a 100-millisecond delay or no stimulus were presented in pseudorandom order. For the startle threshold test, mice were presented with six trial types of varying intensity (no stimulus or 80-, 90-, 100-, 110-or 120-dB pulses -eight presentations of each). Mean startle amplitudes for each condition were averaged.
Fear conditioning Performed essentially as described previously (Powell et al. 2004). Mice were placed in clear plexiglass shock boxes (Med Associates) for 2 min, and then two 90-dB acoustic conditioned stimuli (CS; white noise, each 30 seconds in duration and separated by a 30-second delay) were played. Each CS co-terminated in a 2-second, 0.5-mA foot shock (US). Mice remained in the chamber for 2 min after the second pairing before returning to their home cages. Freezing behavior (motionless except respirations) was monitored at 5-second intervals by an observer blind to the genotype. To test contextual learning 24 h later, mice were returned to the same training context and scored for freezing in the same manner. To assess cue-dependent fear conditioning, mice were placed in a novel environment with an unfamiliar vanilla odor in the afternoon following the contextual test. Freezing was measured first during a 3-min baseline period then during 3 min with the CS playing.

Morris water maze
The MWM and visible platform tests were performed as described previously (Powell et al. 2004) except the probe trial was performed only on day 9. Briefly, a 1.22-m-diameter, white, plastic, circular pool was filled with opaque water (22 ± 1 • C) in a room with prominent extra-maze cues. Mice were trained with four trials per day with an intertrial interval of 1-1.5 min for 8 consecutive days. Mice were placed in one of four starting locations and allowed to swim until they found the submerged platform or until a maximum of 60 seconds had elapsed. Latency to reach the platform, distance traveled to reach the platform, swim speed and percent thigmotaxis (time spent near the wall of the pool) were measured using the Ethovision video-tracking software (Noldus). A probe trial (free swim with the submerged platform removed) was performed on day 9. The percent time spent in the target quadrant and the number of platform location crossings were measured. The visible water maze task was conducted similarly to the traditional MWM with a few changes. A visible cue (black foam cube) was placed on top of the platform. The starting location was held constant, while the platform location was moved for each trial. Mice were given six trials per day for 2 consecutive days, and the latency to reach the visible platform was measured.
Locomotor response to psychotomimetics The apparatus and methods described for the locomotor test were also used to measure locomotion in this task. The test was performed as described by Blundell et al. (2010b). Briefly, at the beginning of each hour of a 3-h locomotor session, mice were given a ∼0.2 ± 0.1 cc intraperitoneal injection of saline (time 0), 0.1 mg/kg MK-801 (1 h) and 0.2 mg/kg MK-801 (2 h). Because fRIM1/KA-cre+ showed increased locomotion to novel environments, they underwent 3 days of this protocol (receiving only saline injections) to habituate before the actual testing on day 4.

Statistics
We used genotype and sex as independent variables for all behavioral tests and performed two-way analyses of variance (ANOVAs) for statistical analysis. Where applicable, time (locomotion tasks), trial (rotarod and visual MWM tasks), day (MWM training) or intensity (PPI and startle threshold) were included as independent variables and a three-way ANOVA with repeated measures was performed. In the MWM probe trial, genotype, sex and location were included as independent variables and a three-way ANOVA was performed. A Student's t-test was used for the planned comparisons analysis in the locomotor tasks and the MWM probe trial. We used Statistica (StatSoft, Inc., Tulsa, OK, USA) for statistical analysis and significance was always taken as P < 0.05.

fRIM1/POMC-cre+ results in RIM1 deletion in DG, arcuate nucleus and select neurons of the cerebellum
We first crossed fRIM1 mice to the POMC-cre transgenic line. This particular POMC-cre line was previously reported to induce recombination selectively in DG granule neurons as well as the arcuate nucleus of the hypothalamus (Balthasar et al. 2004;McHugh et al. 2007). Recombination was reported to begin at 2-3 weeks of age and to remain spatially restricted into adulthood. Assessing the location of RIM1 knockout is particularly challenging due to the absence of antibodies that function selectively in immunohistochemistry (IHC) and the high degree of homology between RIM1 and RIM2. Thus, we first crossed in the previously described conditional Rosa reporter transgenic line (R26R-YFP) (Lagace et al. 2007;Soriano 1999) to obtain fRIM1/POMC-cre+/R26R-YFP mice to label cre recombinase activity in the presence of fRIM1.
IHC staining for YFP in the fRIM1/POMC-cre+/R26R-YFP mice showed successful cre-mediated recombination largely limited to the DG granule neurons, select neurons in the granule and molecular layers of the cerebellum, and arcuate nucleus of the hypothalamus (Figure 1a-c). Co-staining for the neuronal marker NeuN showed that cremediated recombination in DG neurons was robust, but somewhat mosaic. Interneurons in the stratum radiatum, stratum lacunosum-moleculare and stratum lucidum of the hippocampus and in the hilus region did not undergo cremediated recombination, nor did pyramidal neurons in area CA3 or CA1 (Figure 1d-f).
The YFP staining pattern in cerebellum suggested cremediated recombination restricted to neurons of the granular and molecular layers without involvement of purkinje cells (Figure 2a-c). In the molecular layer, consisting largely of interneurons, this was confirmed by co-staining for parvalbumin-containing interneurons showing that almost all YFP-positive neurons in this layer were also positive for the inhibitory interneuron marker parvalbumin, though cremediated recombination did not occur in all parvalbumin positive neurons (Figure 2d-f). The cerebellar granular layer largely consists of granule neurons making mosaic staining for YFP in this region most likely to be in granule cells. To rule out Golgi cells, another common neuron type in the granule cell layer (GL), as responsible for the YFP expression in the GL, we co-stained for the Golgi cell marker neurogranin (Singec et al. 2003). Essentially none of the YFP positive cells co-stained for neurogranin (data not shown), making it most likely that the YFP-positive neurons in this region represent mosaic cre recombination in granule cells of the cerebellum.

fRIM1/KA-cre+ results in RIM1 deletion in CA3 pyramidal neurons as well as mosaic deletion in DG and other brain regions
We next crossed fRIM1 mice to the KA-cre line. This KA-cre line was previously shown to selectively induce recombination in CA3 pyramidal neurons, 10% of DG granule neurons and cerebellar granule neurons, as well as 50% of facial nerve nuclei of the hindbrain (Nakazawa et al. 2002). In area CA3, recombination was first observed at 2 weeks of age with 100% of CA3 pyramidal neurons recombined by 8 weeks. Again, we crossed in the previously described conditional Rosa reporter transgenic line (R26R-YFP; Lagace et al. 2007;Soriano 1999) to generate fRIM1/KA-cre+/R26R-YFP mice, which would label cre recombinase activity in the presence of fRIM1.
Genes, Brain and Behavior (2012) 11: 201-213 IHC staining for YFP in the fRIM1/KA-cre+/R26R-YFP mice showed robust cre-mediated recombination in area CA3 of the hippocampus and mosaic recombination in DG, cortex, thalamus and cerebellum (Figure 1g-i). We did not observe robust staining in the facial nerve nuclei (data not shown). Co-staining for the neuronal marker NeuN showed that cre-mediated recombination in virtually all CA3 pyramidal neurons, hilus region neurons and in a small fraction of DG granule neurons (Figure 1j-l). Interneurons in the stratum radiatum and stratum lacunosum-moleculare of the hippocampus and area CA1 pyramidal neurons did not undergo cre-mediated recombination (Figure 1j-l). We also observed YFP staining in cerebellar purkinje cells and mosaically in neurons of the cerebellar granule layer (Figure 2g-i).

POMC-cre-mediated loss of RIM1 in DG, selected cerebellar neurons and arcuate nucleus leads to increased locomotor responses to the psychotomimetic MK-801
Previous work showed that RIM1α −/− mice had severe behavioral abnormalities that included some schizophreniarelated deficits (Blundell et al. 2010b), learning and memory deficits (Powell et al. 2004), decreased maternal behavior  and social interaction abnormalities (Powell et al. 2004). To test whether loss of RIM1 from the DG, select neurons of the cerebellum and arcuate nucleus of the hypothalamus in the fRIM1/POMC-cre+ mice was sufficient to induce any of the behavioral abnormalities observed in the RIM1α −/− mice, we put the fRIM1/POMC-cre+ mice with sex-matched littermate controls through many of the same behavioral tasks.

Discussion
We have previously shown that global deletion of RIM1α results in multiple behavioral abnormalities, including learning deficits in fear conditioning (Powell et al. 2004) and MWM (Powell et al. 2004), decreased prepulse inhibition (Blundell et al. 2010b), increased locomotor response to novelty (Powell et al. 2004), deficits in social interaction (Blundell et al. 2010b), increased sensitivity to the non-competitive NMDAR antagonist MK-801 (Blundell et al. 2010b) and deficits in maternal behavior . The present results serve to narrow the brain regions responsible for the increased sensitivity to the non-competitive NMDAR antagonist MK-801 and the increased locomotor response to novelty (Table 1). Deficits in maternal behavior and pup rearing were not examined in the present study as they would require specific, complex breeding strategies outside the scope of the present studies.

Characterization of fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice
The fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice lack RIM1 in distinct, partially overlapping brain regions. fRIM1/POMC-cre+ mice exhibit selective cre-mediated recombination in DG granule neurons, arcuate nucleus of the hypothalamus, PV+ GABAergic interneurons of the cerebellar molecular layer (Yu et al. 1996) and scattered neurons of the cerebellar granule layer. This pattern of cre-mediated recombination in fRIM1/POMC-cre+ mice is supported by our qRT-PCR data showing significantly decreased RIM1 mRNA expression levels in both DG and cerebellum, but not in other areas of the hippocampus or the cortex. fRIM1/KA-cre+ mice show robust cre-mediated recombination in area CA3 of the hippocampus and mosaic recombination in hippocampal DG granule neurons, granule and purkinje layers of the cerebellum, thalamus and cortex. Again, the most robust cre-mediated recombination in fRIM1/KA-cre+ mice in CA3 and DG was confirmed by qRT-PCR, while the mosaic, cre-mediated recombination in other parts of the brain was not supported by our qRT-PCR data, suggesting that cre-mediated recombination measured by a reporter transgene does not always accurately reflect cremediated recombination of a different conditional allele. The RIM1 loss in the cerebellum of fRIM1/POMC-cre+ mice and the widespread mosaic expression of cre-mediated recombination in brains from fRIM1/KA-cre+ mice differ from previous reports characterizing the POMC and KAcre driver lines (Balthasar et al. 2004;McHugh et al. 2007;Nakazawa et al. 2002). Changes in expression patterns of Fear conditioning in fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice, respectively. Baseline freezing was measured during the first 2 min in the novel context before tone/footshock delivery. Contextual fear conditioning was measured during a 5-min exposure to the same context the following day. Cued fear conditioning was also measured the following day. Mice were exposed to a different context without the tone for 3 min (pretone), followed by 3 min with the tone playing (tone). (b and d) Social interaction in the open field was measured as the amount of time spent in an interaction zone with either an empty cage (object) or a cage containing a target mouse (social).
cre recombinase due to changes in genetic background, however, have been observed in other cre driver lines (Smith 2011). These findings underscore the importance of determining the regions of cre-mediated recombination in each experimental setting. Although the fRIM1/POMC-cre+ mice and fRIM1/KA-cre+ mice show some overlapping cre expression, the two mouse lines do not share any behavioral abnormalities. This provides evidence that at least some of the abnormal behaviors observed in the RIM1α −/− mice are because of a loss of RIM1 function from one or a few select classes of neurons. Identifying how alterations in presynaptic function in select neuronal subtypes alter select behaviors is important for understanding the cellular basis for these behaviors.

Limited neuronal populations sufficient to induce increased MK-801-induced hyperactivity
Like the RIM1α −/− mice, the fRIM1/POMC-cre+ mice exhibit hypersensitivity to the locomotor enhancing effects of MK-801 (Blundell et al. 2010b). Little is known about the specific brain regions or cell types involved in the locomotor effects of MK-801. Because the fRIM1/POMC-cre+ mice mimic the RIM1 mice, however, our results suggest that RIM1 function in the DG, arcuate nucleus or select neurons of the cerebellum (including PV+ interneurons and granule cells) may modulate the effect of MK-801 on locomotor activity. Although some or all of these regions may in fact play a role in this behavioral phenotype, one interesting possibility is a role for PV+ GABAergic interneurons. Studies on PV+ neurons in the cortex suggest that MK-801 acts preferentially on NMDA receptors expressed on PV+ GABAergic cells leading to decreased inhibition (Beasley & Reynolds 1997;Benes & Berretta 2001;Cochran et al. 2002;Hashimoto et al. 2003;Kinney et al. 2006;Lewis et al. 2005;Li et al. 2002;Xi et al. 2009). Unfortunately, we are unable to further narrow the brain regions or neuronal types involved in this phenotype due to the technical limitation that there are not selective RIM1 antibodies that work for IHC.
While both fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice lead to cre-mediated recombination in the DG, only fRIM1/POMC-cre+ mice exhibit increased sensitivity to MK-801. At first glance, this would appear to rule out a role for loss of RIM1 in DG in this behavioral abnormality. The percentage of dentate granule neurons in which cremediated recombination occurs, however, is clearly much larger in fRIM1/POMC-cre+ mice than in fRIM1/KA-cre+ mice. Thus, we cannot rule out that an effect of loss of RIM1 in a majority of dentate granule neurons (fRIM1/POMC-cre+) rather than a smaller percentage (fRIM1/KA-cre+) may account for the increased sensitivity to MK-801 in fRIM1/POMC-cre+ mice. It is also possible that loss of RIM1 in combinations of the neuronal subtypes identified in fRIM1/POMC-cre+ is leading to the increased susceptibility to MK-801. Finally, though these mice performed normally in the rotarod task, we cannot rule out the possibility that other cerebellar-related abnormalities may become evident if a more extensive coordination test battery were performed.

Lack of memory deficits in fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice
and mosaicism of cre expression in fRIM1/KA-cre+ mice, it is difficult to pinpoint brain regions responsible for the increased locomotion to novelty. One interesting point suggested by our results, however, is the double dissociation between increased locomotion to novelty and enhanced MK-801induced hyperactivity. These two phenotypes do not seem to be mediated by the same set of neurons.

Overview
Overall, both the fRIM1/POMC-cre+ and fRIM1/KA-cre+ mice recapitulate limited components of the phenotypes observed in the RIM1α −/− mice. Both these cre recombinase driver lines have been used in a variety of studies (Banno et al. 2010;Fukushima et al. 2009;Ramadori et al. 2010;Zhang et al. 2009), some of which specifically compared locomotor activity or swim speed in B6 mice with and without the cre transgene and no abnormalities were observed (Tonegawa et al. 2003;Xu et al. 2008). The phenotypes observed in the fRIM1/KA-cre+ mice are subtle and difficult to attribute to RIM1 loss from particular neurons. The hypersensitivity to locomotor activation effects of MK-801 was more narrowly limited to a few cell types and brain regions in the fRIM1/POMC-cre+ mice. One way to better study these brain regions in the future may be to use virally expressed cre recombinase to target specific brain regions without the lack of selectivity that often accompanies cre driver lines.