Mice with mutations of Dock7 have generalized hypopigmentation and white-spotting but show normal neurological function

Abstract

The classical recessive coat color mutation misty (m) arose spontaneously on the DBA/J background and causes generalized hypopigmentation and localized white-spotting in mice, with a lack of pigment on the belly, tail tip, and paws. Here we describe moonlight (mnlt), a second hypopigmentation and white-spotting mutation identified on the C57BL/6J background, which yields a phenotypic copy of m/m coat color traits. We demonstrate that the 2 mutations are allelic. m/m and mnlt/mnlt phenotypes both result from mutations that truncate the dedicator of cytokinesis 7 protein (DOCK7), a widely expressed Rho family guanine nucleotide exchange factor. Although Dock7 is transcribed at high levels in the developing brain and has been implicated in both axon development and myelination by in vitro studies, we find no requirement for DOCK7 in neurobehavioral function in vivo. However, DOCK7 has non-redundant role(s) related to the distribution and function of dermal and follicular melanocytes.

Melanoblasts arise from the neural crest during development, migrate laterally within the embryonic ectoderm, and ultimately differentiate into pigment producing melanocytes found in the skin and hair follicles during postnatal life. Some mutations are known to cause localized hypopigmentation (white-spotting), while others cause generalized hypopigmentation (pigment dilution). White-spotting results from the absence of melanocytes in patches of skin or hair follicles, typically caused by defects in melanoblast ontogeny, proliferation, survival, migration, or differentiation. Anatomically distinct forms of white-spotting include belly spots, dorsal spots, belting of the caudal trunk, piebaldism, head spots, white tipped tail or digits, and peppering (1). Spotting may occur as an isolated phenotype but may be associated with other developmental errors, reflecting the role of neural crest derivatives in many different organ systems. In humans, spotting mutations are sometimes associated with various forms of Waardenburg Syndrome (WS), caused by mutations in many different genes and characterized by localized hypopigmentation, deafness, and other developmental anomalies including defects of hematopoiesis, megacolon, or neurological, cardiac, and craniofacial abnormalities (2).

Pigment dilution occurs when melanocytes are present but are unable to produce or export melanin in normal quantities. This may be due to impaired synthesis of melanin or to defects of melanosomal trafficking and/or exocytosis (3). Various forms of oculocutaneous albinism, including different varieties of Hermansky-Pudlak syndrome (HPS), Chediak-Higashi syndrome (CHS), and Griscelli syndrome result from mutations that cause aberrant melanin synthesis, cargo loading, trafficking, or secretion (4–6). Associated defects of immunological, neurological, or hemostatic function may be observed with such mutations because the formation or intracellular transport of homologous membrane-delimited organelles known as lysosome-related organelles (LROs) may be affected in other cell types. Such organelles include lysosomes, platelet dense granules, and cytotoxic granules of neutrophils, natural killer (NK) cells, and T lymphocytes (7).

Members of the Rho family of GTPases, including Rhos, Racs and Cdc42, are well known for their ability to restructure the actin cytoskeleton and for subsequent effects on multiple biological functions including cell migration, phagocytosis, vesicular transport, apoptosis, and proliferation (8). These factors are implicated in keratinocyte cytophagocytosis of melanocyte dendrites, melanocyte dendrite formation, and melanosome transport and exocytosis, all processes that are important for the deposition of melanin into the skin and growing hair shaft (9–13). Rho GTPases are active when bound to GTP, are inactive in their GDP-bound form, and are tightly regulated by a variety of factors including the guanine nucleotide exchange factors (GEFs) that promote the exchange of GDP for GTP, the GTPase-activating proteins (GAPs) that enhance the GTPase activity of Rho proteins, and the Rho guanine nucleotide-dissociation inhibitors (RhoGDIs) that sequester Rho GTPases in a GDP-bound state (8, 14). GEFs that activate Rho GTPases can be divided into 2 main groups: the classical GEFs containing the nucleotide-exchanging Dbl-homology (DH) domain, and the dedicator of cytokinesis (DOCK)180 superfamily (14).

We now describe moonlight (mnlt), a pigmentation variant that simultaneously displays both pigment dilution and white-spotting phenotypes. The mnlt phenotype consists of an overall dilution of coat color and variable amounts of white-spotting on the belly, tail tip, paws, and genitalia. These pigmentation defects are similar to those found in the classical mutation misty (m), which was discovered nearly 70 years ago (15). The gene underlying the m/m phenotype has not previously been determined. Here, we demonstrate that mnlt and m are allelic and that both phenotypes arise from alterations of Dock7, which encodes a Rho family GEF belonging to the DOCK180 protein family. In cell culture studies, DOCK7 has been shown to activate Racs and Cdc42 and to promote axon formation as well as Schwann cell migration (16, 17). Our findings indicate that DOCK7 plays an important role in pigmentation and may be involved in melanocyte ontogeny and function.

Results

Moonlight Phenotype.

N-ethyl-N-nitrosourea (ENU) mutagenesis of male C57BL/6J mice and subsequent breeding to the third (G3) generation allowed production of mice with homozygous germline mutations. A coat color mutant was identified in the G3 population and shows strict recessive inheritance. The phenotype was named moonlight (mnlt). mnlt/mnlt mice display an overall lightened coat color and hypopigmentation of the distal extremities, particularly the paws, tail tip, and genitalia (Fig. 1). The homozygotes often have belly spots as well. Although spotting is always present in the paws, other types of spotting are incompletely penetrant. It can be scored in the presence of the dominant Agouti allele derived from C3H/HeN mice, used in mapping. Both male and female mnlt/mnlt mice are fertile, but females are poor mothers and their pups invariably require foster care to reach weaning age.

Fig. 1.

Moonlight coat color phenotype. (A) mnlt/mnlt mice have an overall reduction in pigmentation (mnlt/mnlt, Left; C57BL/6J, Right). mnlt/mnlt mice additionally display (B) a white tail tip (mnlt/mnlt, Top; C57BL/6J, Bottom), (C) paws, (D and E) genitalia (D) mnlt/mnlt ; (E) C57BL/6J and (F) belly spot.

Many mutations resulting in hypopigmentation are due to generalized defects in granule function, affecting both melanosomes and other LROs. In particular, the related platelet dense granules may be abnormally formed or secreted, resulting in a bleeding diathesis. Furthermore, immune cells, including NK cells and T cells, use cytotoxic granules for killing target cells. To test LRO function in mnlt/mnlt mice, we assayed bleeding time, but found no defect (Fig. 2). Furthermore, we tested cytotoxic efficiency of NK cells and T cells against target cells using an in vivo assay (Fig. 2). Again, we found no evidence of a generalized LRO anomaly. Further evidence of normal NK cell function comes from the normal resistance of mnlt/mnlt mice to mouse cytomegalovirus (MCMV) in an in vivo test for susceptibility (18). At 10⁵ pfu of MCMV per mouse, signs of sickness were observed in BALB/c mice by day 3 and 50% mortality was observed on day 6, whereas C57BL/6J and mnlt/mnlt mice appeared healthy up to 8 days after inoculation (data not shown). In addition, viral clearance by mnlt/mnlt and C57BL/6J mice was comparable when splenic viral titers were measured at day 5 post infection (data not shown).

Fig. 2.

Moonlight bleeding time and cytotoxicity. Bleeding time for mnlt/mnlt and wild-type C57BL/6J mice. Data shown from one of two experiments with similar results, n = 5 (Left). NK cell cytotoxicity was assessed by in vivo killing of TAP^−/− cells, mnlt/mnlt n = 3, C57BL/6J n = 6 (Middle). Following immunization, in vivo T cell killing was assayed on peptide pulsed target cells, mnlt/mnlt n = 3, C57BL/6J n = 6 (Right). All panels depict means ± SEMs.

Mapping and Identification of the mnlt Mutation.

mnlt/mnlt mice (C57BL/6J background) were outcrossed to C3H/HeN mice and the F1 progeny were backcrossed to the mutant stock or intercrossed. F2 mice were scored based on pigmentation. On 24 meioses the mutation was mapped to chromosome 4 (peak LOD score of 5.4; Fig. 3A). Fine mapping on a total of 57 meioses established a 12.4 Mb critical region delimited by D4Mit301 (88.7 Mb) and D4Mit176 (101.2 Mb). This region overlapped with the previously published critical region for m, a spontaneous mutation that arose on the DBA/J background. The m phenotype is similar to that of mnlt, featuring generalized hypopigmentation, white digits, a belly spot, and white tail tip (15, 19). Although subsequent studies mapped m to chromosome 4 (20, 21), the causative mutation was never identified. To determine whether m and mnlt might be allelic mutations, homozygotes of the 2 strains were crossed. Complementation was not observed, suggesting that mnlt and m represent 2 alleles of the same gene.

Fig. 3.

Identification of the mnlt and m mutations. (A) The mnlt mutation was confined to chromosome 4 using a panel of 128 informative markers. The peak LOD score is 5.4 on chromosome 4. (B) Effect of the mnlt mutation on Dock7 mRNA and protein. Exons 22–28 are deleted, resulting in splicing of exon 21 directly to exon 29. Unbolded text indicates wild-type Dock7 sequence. An aberrant splice joins nt 2596 to nt 3519. Nucleotides of exon 29 are italicized, and the frame-shifted protein product is both highlighted and bolded. (C) Effect of the m mutation on Dock7 mRNA and protein. An anomalous insertion occurs within exon 18. Unbolded text indicates wild-type Dock7 sequence; following the insertion, aberrant nucleotides and the translation product are bolded and highlighted. For exon diagrams in (B) and (C), coding regions are shown in black; non-coding regions in gray.

The mnlt critical region contained a total of 218 annotated genes, derived from an “OR” search of NCBI, OTTMUSG, ENSMUSG, and GENESCAN. Primers were designed for PCR amplification and sequencing of 1483 potential coding exons of genes within this region (22). High-quality coverage of 83.8% of all targeted nucleotide pairs was achieved in amplification products from both wild-type and mnlt/mnlt templates on first pass. No mutations were observed. However, in one candidate gene, Dock7, several contiguous exons could not be amplified or sequenced from the mnlt/mnlt template, although they were readily amplified and sequenced from the C57BL/6J template. We therefore sequenced Dock7 from C57BL/6J and mnlt/mnlt cDNA amplified from the brain and observed a 922 bp deletion in the latter, encompassing bp 2,597 through 3,518 (with reference to the A of the start codon, GenBank accession no. NM_026082.4). Further sequencing at the genomic DNA level defined a 19,025 bp deletion encompassing bases 98,649,759 through 98,668,783 (NCBI mouse assembly version 37.1; supporting information (SI) Fig. S1A). The boundaries of this deletion reside within introns 21 and 28, excising exons 22 through 28 (of 48 total exons). The cDNA sequence is consistent with splicing of exon 21 to 29, which results in a frameshift after amino acid 865, with the incorporation of 30 aberrant amino acids followed by a premature stop codon (Fig. 3B).

The subset of primers specific for Dock7 were subsequently used to amplify and sequence genomic DNA from m/m mice. An LTR retrotransposon insertion after nucleotide 2,045 (with reference to the A of the start codon, GenBank accession no. NM_026082.4) was found to interrupt exon 18 of Dock7 in m/m mice (Fig. S1B). Following amino acid 682 of the DOCK7 protein, ten aberrant amino acids are added, after which chain termination occurs (Fig. 3C).

The 2,130 aa DOCK7 protein is a member of the DOCK180 family, all members of which share 2 conserved domains, DHR-1 (DOCK homology region 1) or CZH-1 (CDM-zizimin homology 1) domain and the more C-terminal DHR-2 or CZH-2 domain. The deletion in mnlt/mnlt mice leads to chain termination after the DHR-1 domain. In m/m mice, chain termination occurs within the DHR-1 domain. The junctional sequences of Dock7^mnlt and Dock7^m alleles have been deposited in GenBank (accession numbers FJ590429 and FJ590430, respectively).

Neurobehavioral Analyses of Dock7^mnlt/mnlt Mice.

In vitro studies have previously suggested that DOCK7 might be important for the formation of axons (23). In addition, knockdown studies have suggested that DOCK7 functions as an intracellular substrate for ErbB2, a plasma membrane tyrosine kinase, and is necessary for in vitro migration of Schwann cells and subsequent myelination of axons (17). Although these reports would suggest an essential role of DOCK7 in development and function of the nervous system, neither Dock7^mnlt/mnlt nor Dock7^m/m mice show obvious neurological impairment. Dock7^mnlt/mnlt mice cannot be distinguished from wild-type animals when measuring depressive behavior in the tail suspension test (Fig. 4A), learning measured by habituation of object exploration (Fig. 4B), working memory using the Y maze test (Fig. 4C), anxiety-like behavior in a light/dark transfer test (Fig. 4D), or locomotor activity (SI Text and Fig. S2). Female Dock7^mnlt/mnlt mice show a trend toward decreased social interaction, although the difference was not significant and males were normal (Fig. 4E). Finally, Dock7^mnlt/mnlt mice are neither blind nor deaf (data not shown). Overall, we find no clear evidence that behavior of Dock7^mnlt/mnlt mice behavior differs from that of wild-type C57BL/6J mice and conclude that neurological function is at least grossly normal in mice with early truncation of DOCK7.

Fig. 4.

Neurobehavioral studies of Dock7^mnlt/mnlt mice. (A) Tail suspension test. Mice are suspended by the tail and monitored for movement. Shown is the cumulative time spent immobile over the course of the trial. (B) Habituation of object exploration. Mice are placed into a field containing 3 plastic objects. Contacts with each object are measured over 5 min of time. Contacts for the 3 objects are averaged for each mouse. Three consecutive trials are performed. (C) Y maze test. The percentage of spontaneous alterations is the number of consecutive triplets of different arm choices out of the total number of arm choices. (D) Light/dark transfer test. Total time spent in the brightly lit compartment was measured (Left) and the total number of transitions between light and dark compartments were counted (Right). (E) Social interaction test. Sociability is measured as the number of entries into a chamber containing a stranger mouse. Differences between Dock7^mnlt/mnlt and C57BL/6J controls are not significant (females P = 0.0619). Social novelty measures the number of entries into a chamber containing a stranger mouse as opposed to a chamber containing a familiar mouse. C57BL/6J is depicted by open bars; Dock7^mnlt/mnlt is depicted by filled bars. Data are represented as means ± SEMs, n = 8.

Discussion

The DOCK180 proteins are guanine nucleotide exchange factors that act on Rho-family GTPases, a class of Ras-homolog small GTPases of which numerous representatives are known in mammals (8, 14). Rho GTPases are known to cycle between GDP-associated and GTP-associated forms. The GTP-associated forms are generally coupled to effector proteins that are stimulated to induce polymerization of the actin cytoskeleton or to maintain tubulin polymerization, which in turn is required for higher-order processes such as migration, proliferation, adhesion, vesicular transport, secretion, maintenance of cell morphology, cytokinesis, ruffling, apoptosis, phagocytosis, polarization, and cell survival (8).

Hydrolysis of GTP to GDP occurs when a GTPase activating protein interacts with the GTP-bound Rho-family GTPase. Regeneration of the GTP-bound form of the protein depends upon the action of a guanine-nucleotide exchange factor (GEF), exemplified by members of the DOCK180 superfamily. To date, this family consists of 11 members that have been divided into 4 subfamilies on the basis of primary sequence alignment: DOCK-A, DOCK-B, DOCK-C, and DOCK-D (24–26). DOCK7 falls within the DOCK-C subfamily and is most closely related to DOCK6 and DOCK8. DOCK family members contain 2 defining domains, DHR1 and DHR2. The DHR1 domain is a lipid binding moiety predicted to allow interaction with phosphatidylinositol (PtdIns) (3, 5)-bisphosphate and PtdIns(3,4,5)P₃ (PIP₃), produced by activated PI3 kinase (PI3K). The DHR1 domain may specify localization of DOCKs within the cell, thus potentially limiting the activity of their target GTPases to specific cellular substructures including the plasma membrane and membranous organelles (16, 26). The DHR2 domain interacts with the Rho GTPase and is required for guanine nucleotide exchange. DOCK7 has been shown to activate Rac and Cdc42 but not RhoA, through its DHR-2 domain (17).

In neuronal cell cultures, DOCK7 plays an important role in axon outgrowth, Schwann cell migration, and axon myelination (16, 17). Rac activation by DOCK7 results in the inactivation of the microtubule destabilizing protein stathmin/Op18 and is necessary for axon outgrowth (16, 27), a process that depends on specific localization of DOCK7 to the developing axon and is likely mediated by the interaction of DOCK7 with PIP₃ (16). Schwann cell migration appears to depend on the binding of axonal neuregulin to ErbB receptors resulting in DOCK7 activation and subsequent activation of both Rac1 and Cdc42 (17). These results, along with the high expression of DOCK7 in the developing brain (16), suggest that DOCK7 may play an important role in neurological functions. DOCK7 has also been found to interact with the tuberous sclerosis (TSC) protein, hamartin (28, 29). Tuberous sclerosis is a multisystem disorder characterized by tumor-like growths, and the TSC protein complex has been shown to have GTPase activity (29).

Mutations in DOCK7 have not previously been reported in any organism. In this paper, we demonstrate that m and mnlt are variant alleles of Dock7 that produce abnormal phenotypes in the homozygous state. Both alleles are predicted to eliminate most of the DOCK7 sequence (68% and 59% of the polypeptide chain, respectively) by causing premature termination and to add extraneous amino acids before termination. Although the mnlt phenotype was identified in an ENU mutagenesis screen, this mutagen generally causes point mutations rather than large deletions. Thus, both Dock7^mnlt and Dock7^m presumably arose as spontaneous mutations. Homozygosity for either allele, or compound heterozygosity for both alleles, causes generalized reduction in pigmentation and white-spotting, suggesting non-redundant function of DOCK7 in two processes related to pigmentation. We propose that the protein may be required for vesicle transport or exocytosis and also for the formation, proliferation, migration, or survival of melanoblasts during embryogenesis. Previous studies have shown that melanoblasts from Dock7^m/m mice are defective in proliferation and differentiation (30), suggesting that DOCK7 plays an important role in some aspect of melanoblast and/or melanocyte function.

DOCK180 family members function in migration, proliferation, and survival of cells (14, 25), and a defect in any of these in melanoblasts could be responsible for the spotting pattern observed. The pigment dilution phenotype suggests a melanocyte defect in the formation or trafficking of melanosomes or their exocytosis to keratinocytes of the hair shaft. The activation of Rac and Cdc42, but not RhoA, by DOCK7 suggests a possible mechanism for the pigmentation phenotypes displayed by Dock7^m/m and Dock7^mnlt/mnlt mice. Both Rho and Rac family members contribute to the formation of melanocyte dendrites that are necessary for the transfer of melanin from melanocytes to adjacent keratinocytes (12, 13). Indeed, numerous studies indicate that dendrite formation in neural cells is controlled by a balance between Rac and Rho activities, with Rho activation promoting neurite retraction and Rac activation promoting neurite outgrowth (13, 31, 32, 33). A similar balance between Rac and Rho activity in dendrite formation has been described in melanocytes, wherein high cAMP levels inhibit Rho activity while promoting Rac activity and dendrite formation. Cdc42 activity has also been implicated in neurite outgrowth, and is suggested to be important for melanocyte function. Additionally, all three of these GTPases appear to play important roles in keratinocyte cytophagocytosis, a process in which the keratinocytes in the skin and hair follicles phagocytose a melanocyte dendrite, leading to a phagolysosome from which melanin granules disperse throughout the cytoplasm.

A number of publications have suggested that DOCK7 might be involved in fundamental neurodevelopmental processes such as axon formation and myelination (16, 17). However, neither Dock7^mnlt nor Dock7^m alleles have a profound effect on nervous system development or function. We cannot formally exclude the possibility that certain subtle aspects of behavior are abnormal in these mice, nor can we be certain that the fragmentary proteins encoded by these alleles are functionally null. However, we tentatively conclude that if DOCK7 contributes to the formation of neural or glial elements of the central or peripheral nervous system, its function as such is redundant. The closely related protein DOCK6 has also been shown to interact with Racs, as well as Cdc42, and to function in neurite growth (35). Moreover, mutations in DOCK8 have been found in patients with mental retardation and developmental disabilities (36). The participation of other GEFs and other Rho family GTPases in neurodevelopment may compensate for DOCK7 deficiency.

Not all of the phenotypic anomalies observed in Dock7^m/m and Dock7^mnlt/mnlt mice are shared. For instance, Dock7^mnlt/mnlt mice do not show the bleeding phenotype seen in Dock7^m/m animals (30). On the other hand, Dock7^mnlt/mnlt females are unable to raise their offspring, while Dock7^m/m mice are adequate mothers. We note that neither of these accessory phenotypes was mapped to the Dock7 locus and therefore could result from other mutations introduced by mutagenesis or spontaneous mutations. Alternatively, the genetic background on which the m and mnlt mutations occurred might be permissive for the hemostatic and nurturing phenotypes, respectively. Finally, cis-acting effects of either of these two mutations (a large insertion in the case of Dock7^m and a large deletion in the case of Dock7^mnlt) might elicit phenotypic differences that are not related to the Dock7 locus itself.

The phenotypes that we have ascribed with confidence to the m and mnlt mutations in Dock7 on the basis of complementation analysis and interstrain comparisons are restricted to abnormalities of pigmentation. We have been unable to detect a defect in NK and T cell effector function in Dock7^mnlt/mnlt mice, which also show normal resistance to MCMV infection. MCMV susceptibility often reveals subtle NK cell defects (37). Additionally, Dock7^mnlt/mnlt mice do not appear to have a bleeding diathesis, which is often caused by defects in platelet dense granules (4, 7). We therefore conclude that if DOCK7 is required for LRO function in mice; this requirement is not widely generalized with respect to cell lineage.

Materials and Methods

Mice.

m/m mice (C57BLKS background) were purchased from Taconic Farms. C57BL/6J and C3H/HeN mice were bred at The Scripps Research Institute. C57BL/6J mice were mutated with ENU as previously described (38, 39) and mnlt/mnlt variants were observed among the G3 population and expanded to form a stock based on their visible phenotype. Mice were maintained under standard housing conditions, and all procedures were approved and performed according to institutional guidelines for animal care. The Dock7^mnlt/mnlt stock was transferred to the Mutant Mouse Regional Resource Centers for distribution (stock no. 030498-UCD).

Mapping and DNA Sequencing.

Homozygous Dock7^mnlt/mnlt males were outcrossed to C3H/HeN females. F1 hybrid mice were subsequently backcrossed to the mutant stock or intercrossed. F2 progeny were phenotyped and linkage was measured using 128 microsatellite markers. The critical region was masked with RepeatMasker, and optimized primers were designed to amplify and sequence all annotated coding regions and splice junctions. Amplification of genomic template from C57BL/6J and mnlt/mnlt mice was performed using two Beckman FX robots, which also cleaned, measured, and diluted the products, adding the correct sequencing primers to each. Sequencing was performed using an ABI 3730XL DNA analyzer. Trace alignments were made using the programs phred and Phrap, and analysis performed with the program consed. High quality coverage was defined as a phred score ≥30 for any bp within the target.

Bleeding Time.

While a mouse is restrained, the tail tip is cut 3 mm from the end. The tail is submerged in saline solution heated to 37 °C. Blood flow is observed while holding the tail down and the time is recorded until bleeding ceases.

Cytotoxicity and MCMV Resistance.

For the in vivo NK cell cytotoxicity assay, splenocyte suspensions from C57BL/6J controls and NK target, TAP1-deficient mice were labeled with a low and high concentration of carboxyfluorescein succinimidyl ester (CFSE) dye, respectively. The two populations were mixed at a 1:1 ratio and a total of 4 × 10⁶ cells were injected i.v. into recipient mice. Recipients were bled two days after injection, and peripheral blood mononuclear cells (PBMCs) were analyzed for CFSE staining by flow cytometry. Killing of targets in TAP1-deficient recipients was used as a control to calculate 0% killing and normalize all measurements.

For the in vivo cytotoxic T lymphocyte (CTL) cytotoxicity assay, mice were immunized i.p with 10⁷ γ-irradiated (1500 Rad) Ad5E1-MEC fibroblasts (murine embryo cells expressing human adenovirus type 5 early region 1) (40). Eight days after immunization, mice were injected with a total of 4 × 10⁶ cells of a 1:1 ratio of naive C57BL/6J splenocytes (CFSE low) and C57BL/6J splenocytes pulsed with an Ad5E1-MEC derived, H2-D^b restricted peptide (VNIRNCCYI; CFSE high). Recipient mice were bled two days after injection, and PBMCs were analyzed for CFSE staining by flow cytometry.

MCMV stock (Smith strain) was prepared and titered as previously described (18, 41). In vivo susceptibility to the MCMV virus was tested as previously described (18). Briefly, MCMV-susceptible BALB/c mice were used as positive controls for susceptibility, while MCMV-resistant C57BL/6J mice were used as negative controls. Mice were inoculated with 10⁵ pfu MCMV via i.p. injection and monitored for signs of illness over 8 days. Viral titers were determined within the spleen at day 5 post infection as previously described (42).

Behavioral Assays.

In the light/dark transfer test, a Plexiglas box is divided into a dark side (10 lux, 14.5 × 27 × 26.5 cm) and a brightly lit side (670 lux, 28.5 × 27 × 26.5 cm). The divider contains a 7.5 × 7.5 cm opening at floor level. Mice are placed in the dark compartment to start. Both the time in light and number of transitions are evaluated for 5 min.

During the Y maze test, mice are placed in a Y maze (3 arms) and recorded by video camera for 5 min. The total number of arm entries and the order of entries are evaluated.

In the habituation of object exploration assay, 3 plastic objects are placed in each of 3 corners of an open field. A mouse is allowed to explore the area for 5 min. Three such sessions are completed for each mouse with a separation of 5 min between sessions. Exploration or contacts are defined as the mouse approaching the object nose-first within 2–4 cm.

Social interaction is assessed in a Plexiglas container that has 3 chambers, each measuring 20 × 40.5 × 22 cm, separated by a clear wall with an opening at floor level. Each outer chamber contains a small, round wire cage (Galaxy Cup, Spectrum Diversified Designs, Inc.). Mice are habituated to the apparatus for 5 min. A stranger mouse of the same sex is placed in one of the wire cages. Sociability is measured as the time spent in the chamber with the stranger mouse over 5 min. For the social novelty test, a second, unfamiliar mouse is placed into the previously empty cage. Time spent in the chamber with the unfamiliar mouse is recorded over 5 min.

For the tail suspension test, each mouse is suspended by its tail using adhesive tape on a metal bar located 30 cm above a flat surface. Immobility is quantified as the duration of time that no whole body movement occurs. Mice are observed for a total of 6 min.