Introduction

In the central nervous system, the neurotransmitter dopamine (DA) is thought to play a pivotal role in the mediation of a diverse array of neurally mediated processes, particularly the regulation of psychomotor behaviour; furthermore, abnormal DAergic neurotransmission has been implicated in a number of pathological conditions such as schizophrenia(Seeman et al. 1984), Parkinson's disease (Hornykiewicz 1966), Huntington's disease (Ginovart et al. 1997) and Gilles de la Tourette's syndrome (Peterson 1996), and in drug-induced movement disorders (Waddington 1989). The functional effects of DA are mediated by two subfamilies of G-protein-coupled receptors: the D1-like family, consisting of D1 and D5 (also known as D1A and D1B) receptors, and the D2-like family, consisting of D2, D3 and D4 receptors (Missale et al. 1998). A lack of ligands with material receptor specificity has stimulated the application of molecular biological techniques to examine the individual in vivo roles of the D1 (Drago et al. 1994; Xu et al. 1994a, 1994b), D2 (Baik et al. 1995; Kelly et al. 1998; Jung et al. 1999), D3 (Accili et al. 1996; Xu et al. 1997), D4 (Rubinstein et al. 1997) and, most recently, the D5 receptor subtype (Holmes et al. 2001)through targeted gene deletion (Drago et al. 1998; Sibley 1999; Waddington et al. 2001). Such mutants offer a powerful opportunity to evaluate the roles of these receptors in DA-mediated behaviours.

It is now well recognised that DAergic regulation of multiple aspects of psychomotor behaviour critically involves both cooperative/synergistic and oppositional interactions between D1-like and D2-like receptor families (Waddington et al. 1994, 1998a). However, lack of ligands with material receptor specificity also precludes resolution of the roles of individual family members in these interactions. The recent introduction of dual knockouts, involving co-deletion of two DA receptor subtypes, offers an important route to addressing this issue. For example, a recent comparative analysis of D2 and D3 single mutants and D2/D3 double mutants revealed that the double mutants develop a motor phenotype which, although qualitatively similar to that of single D2 mutants, is significantly more severe; this suggests that D3 receptors can compensate for some phenotypic aspects of D2 deletion, but that these functional properties of D3 receptors remain masked when the more abundant D2 receptor is expressed (Jung et al. 1999).

There is an emerging body of data to suggest cellular D1-like:D3 interactions (Levavi-Sivan et al. 1998; Ridray et al. 1998; Jung et al. 1999); a role for the D3 receptor in regulating D1-like:D2-like interactions, particularly in terms of a possible inhibitory effect on D1-like:D2 interactions at both electrophysiological and behavioural levels (Xu et al. 1997) has also been proposed. While mice with deletion of the D1 or the D3 receptor have been studied to identify their phenotypic characteristics, the interactive properties of D1 and D3 receptors would be illuminated most powerfully by their co-deletion. To this end, we have generated D1/D3 double mutants. Importantly, in the present study all three mutant lines were generated by the same investigatory team, bred in the same institution and examined for behavioural aspects of phenotype under identical conditions by the same investigators. In particular, as compositing behaviour using automated measures and evaluating behaviour over limited periods can each obscure fundamental consequences of gene deletion (Clifford et al. 1998, 1999, 2000, 2001; Waddington et al. 2001), we included evaluation of phenotype using an ethologically-based, topographical approach to resolve all behaviours in the natural repertoire of the mouse over a prolonged time-frame, from initial exploration through to subsequent habituation.

Materials and methods

Mutant animals

The generation of D1 and D3 receptor mutants was as reported previously (Drago et al. 1994; Accili et al. 1996). Homozygous mutants (D1, D3 and D1/D3) and wild type (WT) control mice were bred in the Department of Medicine, Monash University, and the genotype of each animal determined by Southern blotting; heterozygous founders for both the D1 and D3 genotype were transported from our original colonies at the National Institutes of Health and maintained in a hybrid C57BL/6×129/Sv genetic background (Drago et al. 1994; Accili et al. 1996). For both the D1 and D3 knockout strains, the original C57BL/6 line came from the Jackson Laboratory. In addition, the 129/Sv contribution to the genetic background is from the J1 embryonic stem cell line that was originally used in generating the founding chimeras. The J1 stem cell line was a gift from Dr Rudolf Jaenisch (Whitehead Institute). Within each genotype, breeding pairs were exchanged frequently to maximize background genetic diversity. Mice were housed in groups of three to five under standard animal housing conditions with food and water available ad libitum, and were maintained at 21±1°C on a 12 h/12 h (0900 hours on; 2100 hours off) light/dark schedule. Procedures involving the use of live animals conformed to the Australian National Health and Medical Research Council code of practice.

Topographical assessment of behaviour

On experimental days, mice were removed from their home cage and placed individually in clear glass observation cages (36×20×20 cm). Behavioural assessments were carried out in a manner similar to that described previously (Clifford et al. 1998, 1999, 2000, 2001; Ross et al. 2000) using a rapid time-sampling behavioural checklist technique. For this procedure, each of ten randomly allocated mice was observed individually for 5-s periods at 1-min intervals over 15 consecutive minutes, using an ethologically based behavioural checklist. This allowed the presence or absence of the following individual behaviours (occurring alone or in any combination) to be determined in each 5-s period: sniffing; locomotion (co-ordinated movement of all four limbs producing a change in location); total rearing (rearing of any form); rearing seated (front paws reaching upwards with hind limbs on floor in sitting position); rearing free (front paws reaching upwards away from any cage wall while standing on hind limbs); rearing to wall (front paws reaching upwards onto or towards a cage wall while standing on hind limbs); sifting (sifting movements of the front paws through cage bedding material); grooming (of any form); chewing (chewing movements directed onto cage bedding and/or fecal pellets without consumption); eating (chewing with consumption); climbing (jumping onto cage top with climbing along grill in inverted or hanging position); stillness (motionless, with no behaviour evident). This cycle of assessment by behavioural checklist over a 15-min period (0–15 min) was repeated twice (20-35 min and 40–55 min); thereafter, eight additional cycles of otherwise identical assessments were repeated at 80–90, 120–130, 160–170, 200–210, 240–250, 280–290, 340–350 and 360–370 min. All assessments were made by an observer who was unaware of the genotype of each animal.

Data analysis

For determination of ethograms over the phase of initial exploratory activity, total "counts" for each individual behaviour were determined as the number of 5-s observation windows in which a given behaviour was evident, summed over the initial 3×15 min (0–15, 20–35, 40–55 min) cycle periods and expressed as means±SEM; this was to consolidate data over a 1-h period of exploration so as to facilitate general comparisons with other studies in the field, which have commonly accumulated data over similar periods, and to allow specific comparisons with our previous work using these methods. These counts were analysed using ANOVA followed by Student's t-test or, in instances where data distribution deviated from normality, using the Kruskal-Wallis non-parametric ANOVA followed by Mann-Whitney U-test. For determination of habituation profiles for these ethograms, total counts for each individual behaviour were summed as above over each of the following periods: 0–10, 20–30, 40–50, 80–90, 120–130, 160–170, 200–210, 240–250, 280–290, 340–350 and 360–370 min; these were expressed also as means±SEM. Data were analysed using repeated-measures ANOVA, following square-root transformation to allow examination of interaction effects in the absence of non-parametric techniques for interaction terms (Clifford et al. 1999, 2000, 2001; Ross et al. 2000).

Results

General parameters

Relative to WT controls (n=20:8 females, 12 males; weight 26±1 g, age 102±4 days), D1 mutants (n=19:8 females, 11 males; weight 17±1 g, age 110±9 days) showed a significant reduction in weight (−40%, P<0.05); in contrast, D3 mutants (n=23:10 females, 13 males; weight 29±1 g, age 105±4 days) showed no difference in weight; D1/D3 double mutants (n=21:10 females, 11 males; weight 18±1 g; age 97±4 days) showed a significant reduction in weight (−35%, P<0.05). There were no gross neurological deficits in any mutant line on qualitative inspection of posture, reactivity to handling and general activity. Furthermore, no neurological deficits were detected on formal neurological examinations conducted on each line, the protocol of which is described in Drago et al. 1994.

Ethogram over exploratory phase

Over an initial 1-h period of exploration, D1 mutants were characterised as follows: increases in sniffing (P<0.001) and locomotion (P<0.001), with reductions in rearing free (P<0.001), rearing seated (P<0.01), grooming (P<0.01), chewing (P<0.001) and stillness (P<0.05); there were no significant differences in sifting, total rearing, rearing to wall or climbing (Fig. 1). In contrast, D3 mutants were characterised as follows: increases in sniffing (P<0.001), locomotion (P<0.001), total rearing (P<0.001) rearing free (P<0.001) and rearing to wall (P<0.001), with reductions in grooming (P<0.01), rearing seated (P<0.01) and stillness (P<0.05); there were no significant differences in sifting and chewing. Behaviour of D1/D3 mutants over the same time frame evidenced a profile similar to that of D1 mutants; thus D1/D3 mutants were characterised as follows: increases in sniffing (P<0.001) and locomotion (P<0.001), with reductions in rearing free (P<0.001), rearing seated (P<0.01), grooming (P<0.01), chewing (P<0.001) and stillness (P<0.05); there were no significant differences in sifting, total rearing, rearing to wall and climbing, with only low levels of intense grooming and eating noted in each genotype (Fig. 1).

Fig. 1.
figure 1

Behavioural counts for sniffing, locomotion, total rearing, rearing free, rearing seated, rearing to wall, grooming, sifting, chewing, eating, climbing and stillness in wild type, D1, D3 and D1/D3 double mutant mice. Data are mean counts±SEM over a 1-h period of initial exploration for each group of animals. ***P<0.001, **P<0.01, *P<0.05 versus wild type (n=20 wild type, n=19 D1 mutant, n=23 D3 mutant and n=21 D1/D3 double mutant mice per group)

Full size image

Ethogram over habituation phase

Sniffing declined over time in all genotypes, though habituation was delayed, to similar extents, for each mutant line relative to WT (time×genotype interactions: D1, P<0.001; D3, P<0.01; D1/D3, P<0.001; Fig. 2). Though locomotion also declined over time for each genotype, habituation was delayed, to similar extents, for D1 and D1/D3 (time×genotype interactions: D1, P<0.01; D1/D3, P<0.01), whereas initially increased locomotion in D3 mutants declined to the level of WT. Total rearing, rearing seated and rearing to wall evidenced a profile characterised by some delay in habituation for each mutant line (time×genotype interactions: total rearing: D1, P<0.001, D3, P<0.001, D1/D3, P<0.001; rearing seated: D1, P<0.001; D3, P<0.05, D1/D3, P<0.001; rearing to wall: D1, P<0.001; D3, P<0.01, D1/D3, P<0.05); conversely, while D1 and D1/D3 mutants evidenced low levels of rearing free throughout habituation, initial increases in this behaviour observed in D3 mutants readily declined to the level of WT (time×genotype interaction, P<0.01). Although grooming was initially reduced in D1 mutants during the exploratory phase, this behaviour subsequently occurred to excess over the later phases of habituation (time×genotype interaction, P<0.01); conversely, D3 and D1/D3 mutants did not show significant changes in grooming over the habituation phase. Sifting evidenced habituation to generally similar extents for each genotype. There were no overall effects of gender on any topography of behaviour, nor were there any genotype×gender interactions.

Fig. 2.
figure 2

Behavioural counts for sniffing, locomotion, total rearing, rearing seated, rearing free, rearing to wall, grooming and sifting in wild type, D1, D3 and D1/D3 double mutant mice. Data are mean counts±SEM per 10-min period at indicated intervals over a 370-min period of habituation for each group of animals (n=20 wild type, n=19 D1 mutant, n=23 D3 mutant and n=21 D1/D3 double mutant mice per group)

Full size image

Discussion

Dominant expression profiles for D1 and D2 receptors are generally consistent with the functional prepotence ascribed to these subtypes, such that independent functions for remaining members of the D1-like (D5) and D2-like (D3 and D4) families have been less clearly delineated (Missale et al. 1998; Sibley 1999; Waddington et al. 2001). Also, while D1-like:D2-like interactions putatively involving D1 and D2 receptors are well established (Waddington et al. 1994, 2001), it is possible that D5, D3, or D4 receptors exert effects through direct modulation of or interaction with D1 and/or D2 receptor function. This study describes the behavioural phenotype of D1, D3 and D1/D3 double mutant mice as an important route to investigating such modulation/interaction.

During the exploratory phase, D1 mutants evidenced an ethogram characterised by alterations in a number of behavioural topographies, including increases in sniffing and locomotion with reductions in rearing free, rearing seated and chewing. Increases in "activity", as operationalised in terms of line-crossings or assessed automatically in photobeam cages, have been noted previously on examining D1 mutants from an independently generated line having a different construct design (Xu et al. 1994a, 1994b); similarly, 3-week-old mutant mice with targeted expression of an attenuated diphtheria toxin gene to D1 positive neurones (Wong et al. 2000) showed increased "activity". However, the present data illustrate further how compositing spontaneous behaviour into otherwise undifferentiated activity can obscure phenotypic effects (Clifford et al. 1998; Waddington et al. 2001). Specifically, the present ethologically based approach reveals a considerably broader phenotype at the level of individual topographies of exploratory behaviour; this includes not just increases in ethologically defined locomotion but also subtle shifts in individual elements of rearing. Thereafter, D1 mutants evidenced delayed habituation in a number of behavioural topographies, such that increases found over initial exploration became accentuated; this elaborates our previous finding (Clifford et al. 1998) that confining assessments to a limited time-frame can also obscure important phenotypic effects of gene deletion, as in D1 mutants there appears to be disruption to the neuronal processes which sculpts the changing topography of behaviour from initial exploration, through habituation, to quiescence (Waddington et al. 2001).

Some aspects of phenotype showed apparent differences from our initial report on this D1 mutant line (Clifford et al. 1998, using the D1A nomenclature), primarily a greater increase in locomotion, increased sniffing, and decreased rearing seated, grooming and stillness. On the basis of response competition considerations (Waddington et al. 2001), these differences are likely to derive principally from the greater increase in locomotion, which is usually accompanied by sniffing but is incompatible with sedentary rearing, grooming and stillness. Shifts in time course also appear to be relevant. Though we reported previously (Clifford et al. 1998) increased but here decreased grooming, this decrease was confined to the initial exploratory phase; thereafter, habituation of this behaviour was delayed in D1 mutant mice such that grooming did occur to excess subsequently. As mice used in this study were on a hybrid C57BL/6 and 129/Sv genetic background, independent strain-specific modifier genes on a mixed genetic background may have contributed to these phenotypic differences within the "same" D1 mutant line; this would substantiate previous concerns among laboratories breeding mutant lines from small numbers of progenitor mice, and indicates further the need for congenic mutant lines in behavioural experiments (Waddington et al. 2001; McNamara et al. 2002a, 2002b; Tomiyama et al. 2002). Also, there are recognised to be poorly understood differences in behavioural findings between different laboratories using the "same" methodology (Crabbe et al. 1999); thus, it may be relevant that our initial studies were conducted in Ireland (Clifford et al. 1998) while the present work was conducted "similarly" in Australia. Such considerations may account also for certain differences in levels of behaviour in wild type mice. This does not detract from intra-laboratory comparative studies such as these, but makes it more problematic to compare studies on an inter-laboratory basis.

During the exploratory phase, D3 mutants evidenced a distinct ethogram characterised by alterations in a number of behavioural topographies, including increases in sniffing, locomotion, total rearing, rearing free and rearing to wall, with reductions in rearing seated and grooming. An early increase in activity and undifferentiated rearing events, as assessed automatically in photobeam cages or operationalised as line-crossings over an open field, has been noted previously on examining D3 mutants from the present line (Accili et al. 1996; Steiner et al. 1998), and from an independently generated D3 mutant line having a different construct design (Xu et al. 1997). A third line of D3 mutants on a mixed genetic background evidenced no increase in activity in terms of photobeam interruptions over a 30-min period, while mice of the present D3 mutant line back-crossed for an additional generation into C57BL/6 have also been reported not to evidence "hyperactivity" in terms of photobeam interruptions (Boulay et al. 1999). Thereafter, D3 mutants did not evidence the prominence of delayed habituation in several topographies of behaviour that characterised their D1 mutant counterparts. These findings indicate a distinct but more subtle phenotype in D3 mutant lines that may be influenced by independent strain-specific modifier genes in relation to genetic background for the "same" knockout (McNamara et al. 2002b).

In the control of DA-mediated behaviours, D1-like and D2-like receptors do not function independently, but rather are subject to critical D1-like:D2-like interactions. Typical D2 -like-initiated behaviours, such as stereotyped sniffing and locomotion, are regulated in a co-operative/synergistic manner by tonic or phasic activity through D1-like receptors; conversely, atypical behaviours appear to be regulated in an oppositional manner, such that vacuous chewing has its genesis in tonic or phasic activity through D1-like receptors but is released/enhanced by reduction in DAergic activity through D2-like receptors (Waddington et al. 1994, 1998a, 1998b). However, the involvement of individual family members in these effects is poorly understood. There are some emergent data to suggest both cellular D1-like:D3 interactions (Peterson 1996; Levavi-Sivan et al. 1998; Ridray et al. 1998; Jung et al. 1999) and a role for the D3 receptor in regulating D1-like:D2-like interactions, particularly in terms of a possible inhibitory effect on D1-like:D2 interactions at both electrophysiological and behavioural levels; thus, behavioural evidence has been offered that the D3 receptor may inhibit co-operative /synergistic D1-like:D2-like regulation of activity in terms of otherwise undifferentiated photobeam interruptions(Xu et al. 1997). In the face of distinct phenotypic ethograms for the D1 and D3 mutants studied here, the topographical profile of D1/D3 double mutants over both exploratory and habituation phases was essentially indistinguishable from that of their D1 counterparts. The only exception to this general observation relates to aspects of grooming behaviour (see above). Thus, it appears that absence of D1 receptors dominates the phenotype of D1/D3 double mutant mice. As absence of D3 receptors did not interact with these functional consequences of D1 receptor ablation, no evidence for an effect of D3 receptors on D1-dependent behavioural topographies or on how tonic DAergic activity through D2 (or through D5 or D4) receptors might influence such topographies was apparent.

Recently, another laboratory (Karasinska et al. 2000) has reported phenotypic comparisons between D1, D3 and D1/D3 double mutants. They reported: no alteration in undifferentiated activity in terms of photobeam interruptions over a 1-h period, for any genotype; reduction in line crossings that was more prominent in D1/D3 than in D1 mutants, but unaltered in D3 mutants; reduction in undifferentiated rearing events that was more prominent in D1/D3 than in D1 mutants, while such rearing events were increased in D3 mutants; in the plus-maze, the only finding was a greater number of open arm entries in D3 mutants. However, while these studies utilised the present D1 mutant line, they involved an alternatively sourced D3 mutant line, with D1/D3 double mutants derived using a breeding strategy distinct from that adopted here, and only male mice were used; furthermore, the behavioural approach adopted differed radically from our own ethologically based, topographical approach. Though they reported some aspects of phenotype in D1/D3 double mutants to be more similar to D1 than to D3 mutants, so many differences in experimentation, each capable of influencing apparent phenotype (Crabbe et al. 1999; Waddington et al. 2001), make it difficult to compare the present results directly with these findings.

In summary, the behavioural ethogram described for D1 and D3 mutant mice is broadly similar to results obtained using comparable methodology on the same lines. The phenotype of the D1/D3 double mutant line was largely reflective of the D1 mutant component. Future drug studies with selective D1- and D2-like agents in D1/D3 double mutants should provide important additional insights into the role of the D3 receptor in the modulation of behaviour under phasic challenge.