Introduction

Changes in intracellular Ca2+ levels control a variety of cellular functions, including neurotransmitter release from synaptic vesicles [1–3]. This process involves the fusion of lipid bilayers between a vesicle and the plasma membrane. The soluble N-ethylmaleimide sensitive factor attachment receptor (SNARE) proteins are a large family of membrane-associated proteins that are critical for synaptic vesicle fusion [4–6]. The heterotrimeric SNARE complex is formed by syntaxin 1, synaptosomal-associated protein 25 kDa (SNAP25), and vesicle-associated membrane protein (VAMP)/synaptobrevin [7]. Botulinum and tetanus toxins cleave VAMP1 and VAMP2 to inhibit neurotransmitter release, defining the important role of this complex during vesicle fusion [8–12]. However, the SNAREs are unlikely to represent a minimal fusion complex as multiple proteins are involved in this cellular event. The SNARE proteins are membrane embedded and form transcomplexes, bringing the vesicle in close proximity to the plasma membrane [13, 14]. Early evidence demonstrated that changes in intracellular Ca2+ levels regulate the process of vesicle fusion; however, the SNARE complex does not bind Ca2+ directly. Ca2+ was shown to trigger SNARE complex formation in pheochromocytoma 12 cells [15]. The results of several studies have suggested that synaptotagmin [16–18] and calmodulin [19] are the Ca2+ sensors that regulate vesicle fusion through the SNARE complex. Calmodulin was shown to directly interact with VAMP2 and regulate exocytosis [20–22]. The calmodulin interaction domain was localized to amino acids 77–94 of VAMP2, downstream of the tetanus toxin cleavage site [23].

In the mouse, there are six VAMPs (also known as synaptobrevin) and an additional seven proteins that have a synaptobrevin domain. Knockout mutants have been produced for VAMP2, VAMP3, and VAMP8. VAMP2 mutants die immediately after birth and show no detectable histopathology. Neurons in the affected mice show a drastic reduction in Ca2+-triggered vesicle fusion [24]. Consistent with the severity of the mutant phenotype, VAMP2 is extensively expressed in the central nervous system (CNS; Gene Expression Nervous System Atlas [GENSAT] image no. 47595). VAMP8 knockout mice develop normally but have pancreatic abnormalities. The affected mice have severe defects in the exocytosis of zymogen granules from acinar cells [25]. The VAMP3 knockout mouse had no detectable mutant phenotype despite extensive evaluation [26], suggesting the existence of functional redundancy. None of these knockouts exhibit any sort of developmental defect, indicating that the VAMP genes are not required for embryonic development. Together, these in vivo models have been useful for functional studies of the VAMPs and have shaped the present understanding of SNARE-mediated vesicle fusion. Mutations in the other VAMP genes have yet to be reported; however, the phenotype and underlying molecular mechanisms of pathology should be predictable based on these data and on the knowledge of the expression pattern of the gene family. The expression pattern of VAMP1 in the rat brain is restricted [27], which suggests that the phenotype of a mouse with a VAMP1 mutation would also be limited.

In this report, we describe the identification of a nonsense mutation in VAMP1/synaptobrevin associated with the lethal-wasting mutant phenotype. The lethal-wasting (lew) mutant spontaneously arose at the Jackson laboratory in 1985 on a C3H/HeSnJ background. Lethal wasting is inherited in an autosomal recessive manner and manifests as a neurological phenotype. At 2 weeks postpartum, the mutant displays neurological signs such as curling its front paws inward and clasping its back legs tightly together. The affected mice can move their limbs although not in any purposeful manner. Near P12, the mutants are noticeably immobile, lay on their side, and die preweaning. The gross histopathological examination of the brain showed no lesions. This disorder was linked to a relatively large region on chromosome 6 [28]. A positional cloning strategy was used to identify this mutation; however, rather than extensive breeding to genetically narrow the region, we used informatics resources to select candidate genes from a relatively large genetic region. In this manuscript, we report the identification of a nonsense mutation in VAMP1 that is associated with the lethal-wasting phenotype. The mutant is a true null, as VAMP1 is not detected in mutant tissues. The expression pattern of VAMP1 in the mouse suggests a limited but vital group of neurons that are dysfunctional in the lethal-wasting mutant.

Materials and methods

Genotyping and husbandry

All animals were bred and maintained under standard conditions at the University of Nebraska Medical Center (UNMC) research vivarium in accordance with the Institutional Animal Care and Use Committees standards. A mating cross of C3H/HeDiSnJlew×CAST/EiJ was established to generate F1 mice for an intercross progeny testing. The matings that produced affected offspring were maintained as a mapping cross to generate F2 mice for genotyping. The offspring were counted and visually observed daily until weaning age to determine if perinatal or in utero lethality occurred. Deoxyribonucleic acid (DNA) was isolated from tail tips excised from 10-day-old F2 pups. Approximately one eighth of an inch of tail was excised and then incubated overnight at 55°C in 650 μl of tail digestion buffer (50 mM Tris, 400 mM NaCl, 100 mM ethylenediamine tetraacetic acid, 0.5% sodium dodecyl sulfate (SDS), and 400 μg proteinase K). After the tail tip was fully digested, 175 μl of 6 M NaCl was added and mixed thoroughly. Samples were centrifuged at 14,000 × g for 20 min. Supernatant was recovered and added to an equal volume of 100% ethanol. The DNA was pelleted by centrifugation at 14,000 × g for 10 min, dried, and resuspended in ultrapure H2O to a final concentration of 40 ng/μl.

The offspring were genotyped with several short tandem repeat polymorphism markers in the disease region using a standard PCR protocol. Briefly, DNA was amplified by PCR using the following conditions: 40 ng of template DNA in a 10 μl PCR reaction mixture containing 1.0 μl PCR buffer (100 mM Tris-HCl [pH 8.8], 500 mM KCl, 15 mM MgCl2, and 0.01% w/v gelatin); 200 μM each deoxyadenosine triphosphate, deoxycytidine triphosphate, deoxyguanosine triphosphate, and deoxythimidine triphosphate; 2.5 pmol of each forward and reverse primer; and 0.25 U Taq polymerase. Reaction mixtures were subjected to 40 cycles of 94°C for 30 s, 53°C for 60 s, and 72°C for 60 s. Amplification products were electrophoresed at 100 V for 1 h on a 4% agarose gel containing 0.5 μg/ml ethidium bromide and visualized by ultraviolet transillumination. C3H/HeDiSnJ and CAST/EiJ control DNAs were included in each set of reactions for proper sizing of strain-specific alleles to ensure proper genotyping.

Expression analysis

Quantitative RT–PCR was used to determine relative expression levels of candidate genes. Total brain ribonucleic acid (RNA) was isolated from four mutant and control littermates using Trizol reagent (Invitrogen) per manufacturer’s instructions. The RNA was DNase treated with DNA-free (Ambion) to remove the contaminating DNA, per manufacturer’s instructions. First, strand synthesis was performed using the RETROscript kit (Ambion) on 2 μg of RNA template, per manufacturer’s instructions. Reaction products were diluted 1 to 100, and 1 μl was used as template for amplification using the following 10-μl-reaction mixture: 100 pmol each forward and reverse primer, and 6 μl Sybr Green PCR master mix (ABI). Forward and reverse primers were chosen for each gene using Primer Express 3.0 (ABI). Quantitative PCR was carried out on an ABI 7500 using the default cycling parameters. Reactions were performed in triplicate for the four separate RNA samples for wild type and mutant. As a measure of amplification, the number of cycles to a manually set ΔR n (dye fluorescence) threshold was determined as the ΔC t value (ABI 7500 system software). ΔC t values were obtained for each sample, a total of 12 for each gene (four triplicated replicates). The average ΔC t for each gene was calculated for each data set, as well as the standard deviation and standard error. Significance was determined by the t-test. The ΔΔC t was derived by comparison to a control gene, β-actin, for each sample. Comparisons between mutant and wild-type control samples were made for fold-change estimation using the comparative C t method, \(2^{{ - \Delta \Delta C_{t} }} \), where ΔΔC t = ΔC t test gene-ΔC tβ-actin; relative expression was calculated per 1,000 molecules of β-actin, \({\text{1,000/2}}^{{ - \Delta \Delta C_{t} }} \).

Mutation analysis

Mutation analysis was performed by directly sequencing RT–PCR products amplified from RNA and confirmed by sequencing PCR products amplified from genomic DNA. RT–PCR primers were designed from exon sequences of VAMP1 to amplify several fragments across the coding region. RT–PCR was carried out on mutant and wild-type RNA from multiple tissues using the Titan one-tube RT–PCR system (Roche). Amplified products were recovered and sequenced as stated below. For confirmation of the mutation, we sequenced genomic DNA using primers that were designed flanking the exons of VAMP1. DNA was amplified as before. For the strain survey, DNA samples from several strains were purchased from the Jackson Laboratory mouse DNA resource. The appropriate bands were excised, and DNA was recovered using the Qiagen gel extraction kit, per manufacturer’s instructions. The recovered fragments were sequenced by the DNA-sequencing core at UNMC. Sequence analysis was performed using Sequencher 4.5 (gene codes). Protein sequence analysis was performed using Clustal [29].

Histology

Tissues were fixed via transcardial perfusion of the mouse. Tissues were removed from mutant and age-matched 14-day-old littermates and fixed in 3.7% paraformaldehyde. Tissues were paraffin embedded by the Eppley Center Histology Core. Tissues were sectioned in the sagittal plane at 8 μm thick on a Lieca RM2135 microtome. Sections were mounted on poly-l-lysine-coated slides. Hematoxylin and eosin staining was performed as before for general histological analysis of the tissue [30]. Sections were visualized under bright-field on a Ziess AxioPlan 2 and imaged with the Ziess AxioCam using Axiovision 4.3 software.

Western blot analysis

The Western blot analysis was performed using 20 μg of protein per sample. A control and a mutant brain were dissected from postnatal day 14 (P14) mice and homogenized in radioimmunoprecipitation buffer (1 × Tris buffered saline, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 0.04% sodium azide, 1 mM phenylmethylsulphonylflouride, 1 × complete protease inhibitor cocktail [Roche]) using an Ika ultra-turrax T8 homogenizer at 4,000 rpm. Samples were centrifuged at 14,000 × g for 20 min at 4°C to remove debris. The supernant was boiled for 10 min in NuPAGE lithium dodecyl sulfate sample buffer (Invitrogen) and electrophoresed at 200 V for 30 min on 12% Tris-Bis NuPAGE gels (Invitrogen). Gels were equilibrated in 1 × NuPAGE transfer buffer (Invitrogen). Proteins were transferred onto equilibrated polyvinylidenfluoride membranes using the X-Cell II Blot Module (Invitrogen), per manufacturer’s instructions. Transfer was performed at 30 V for 2 h at room temperature in transfer buffer. After transfer, blots were washed in phosphate-buffered saline (PBS) and then incubated in Odyssey blocking solution (LiCor Technologies) for 1 h at room temperature. Primary antibody was diluted in Odyssey blocking solution, anti-VAMP1 at 1:3,000 (Abcam) or anti-β-actin at 1:1,000 (Santa Cruz), and incubated on the membrane overnight at 4°C. The blot was then washed in three changes of PBS. The appropriate secondary antibody was incubated on the membrane at 1:5,000 in Odyssey blocking solution for 1 h at room temperature (AlexaFluor680 [Molecular Probes] or IRDye800CW [Rockland]). Images were detected with the Odyssey infrared imaging system (LiCor Technologies) using the proper infrared channels and visualized using Odyssey v.1.2 software.

Immunohistochemistry

Immunohistochemistry was performed as before [31] using an antibody specific to VAMP1 (Abcam). Control and mutant P14 mice were fixed via transcardial perfusion, and brains were dissected. Tissues were prepared and sectioned as before. Slides were treated in the following washes to deparaffinize and hydrate the sections: xylene, 10 min, two changes; 100% ethanol, 5 min, two changes; 95% ethanol, 5 min; 70% ethanol, 3 min; PBS, 5 min. Tissue sections were blocked in 1 × PBS, 0.1% Tween, 5% goat serum for 1 h, then incubated with primary antibody (1:1,000) in blocking solution overnight at 4°C in a humidified chamber. The slides were washed with three changes of PBS for 5 min each and then incubated with secondary antibody, AlexaFlor 488 (Molecular Probes), 1:200 in blocking solution. Slides were washed in three changes of PBS for 5 min each and briefly counterstained with 4′,6-diamidino-2-phenylindole (Sigma) 1:200 in PBS, washed in PBS again, and mounted with Vectashield (Vector Laboratories). Sections were visualized by florescent microscopy on a Ziess AxioPlan 2 and imaged with the Ziess AxioCam using Axiovision 4.3 software.

Results

Characterization of the lethal-wasting phenotype

Lethal-wasting is an autosomal recessive mutation. In the mapping cross, we observed 25.6% affected mice with no indication of in utero loss or immediate death after birth. The affected mice die near P15. By P10, the mutants are noticeably immobile and lay on their side. The mutants fail to attempt to right when placed on their backs; otherwise, in earlier perinatal stages, the mutants are of normal size and difficult to discern from their normal littermates. The affected mice can move their limbs although not in any purposeful manner. Hematoxylin and eosin were used to stain brain sections. No gross histological defects were observed (data not shown).

Genetic mapping of the lethal-wasting mutation

The lethal-wasting disease interval was reported to be between markers D6Mit55 and D6Mit111, a distance of approximately 20 mb, containing more than 200 genes. We created a C3H/HeDiSnJlew×CAST/EiJ cross to generate a total of 256 F2 offspring for genetic fine mapping. The disease region was narrowed to approximately 1 mb by mapping recombination events that had occurred in five F2 offspring (Fig. 1). Recombination events had occurred in three unaffected mice that were progeny tested to ascertain carrier status. Mouse 127 never had affected offspring when mated to an obligate heterozygote (0 out of 27 pups), whereas mouse 242 did have affected offspring. Their recombination events both mapped between markers D6Mit193 and D6Mit24, thus excluding everything proximal to D6Mit193. Mouse 252 was unaffected and homozygous for the C3H/HeDiSnJ chromosome at markers D6Jpk2 and D6Mit334, thus excluding everything distal to these markers. Three markers remained completely linked with the disease phenotype. The Ensembl genome browser was used to identify the genes between markers D6Mit193 and D6Jpk2. Several informatics tools were used to evaluate all 41 of the genes within the interval. Nineteen of the genes had been previously associated with a disparate mutant phenotype in the mouse. Of the remaining genes, five were chosen for mutation screening based on a predominantly neuronal expression pattern and putative function (Table 1).

Fig. 1
figure 1

Genetic map of the lethal-wasting region on chromosome 6. On the left, the physical map of chromosome 6 is shown with the genetic markers that were used for fine mapping. On the right, five mice that define the critical genetic interval for the disease are shown. Mapped recombination events are depicted as colored chromosomes, the wild type CAST/EiJ chromosome (black) and mutation carrying C3Hlew chromosome (red). The mouse identification number and its phenotype are given (affected [aff] and unaffected [wt])

Full size image
Table 1 Candidate gene analysis for genes in the lethal-wasting interval. Five genes were selected based on the expression data from UniGene’s EST profile viewer and putative function. Quantitative RT–PCR was used to identify significantly decreased levels of Vamp1 expression in lethal-wasting mutant tissue compared to wild-type littermates
Full size table

Identification of a nonsense mutation in VAMP1

It is generally thought that approximately 30% of known mutations in the mouse results in a disruption of the expression of the gene. Thus, an efficient first-pass mutation screen was performed on five genes using quantitative RT–PCR (qRT–PCR) to detect differences in their relative expression between mutant and wild-type tissues. Total RNA was isolated from four mutant and four wild-type brains and used as template for qRT–PCR. Five genes were evaluated: atrophin 1, VAMP1, enolase 2, guanine nucleotide-binding protein, beta 3, and G protein-coupled receptor 162. β-Actin was used as a control and a benchmark for the determination of fold change between mutant and wild-type message levels. VAMP1 mRNA levels were significantly decreased (p < 0.001) in mutant brains compared to wild-type littermates (Table 1). None of the other genes showed any significant differences in their expression levels between mutant and wild-type RNA samples.

VAMP1 was amplified by the RT–PCR from the mutant and control brain RNA and sequenced. A G190T transversion resulting in a premature stop codon was identified in the mutant (Fig. 2a). This mutation was confirmed by sequencing DNA, amplified by the PCR from the heterozygous and affected mice. This mutation was not detected in a strain survey of 20 common laboratory mouse strains, including the strain of origin (data not shown).

Fig. 2
figure 2

Identification of a nonsense mutation in Vamp1 in the lethal-wasting mutant. Sequence chromatograms from wild-type (panel a), heterozygous (panel b), and mutant (panel c) show the G190T transversion (highlighted in blue). This mutation truncates VAMP1 upstream of functional domains (panel d). VAMP1 amino acid sequences are shown from several species: mouse, human, frog, rat, cow, and opossum. VAMP2 and VAMP3 sequences from the mouse are also shown. Shaded areas show blocks of high conservation. The position of the nonsense mutation is noted (lewvamp1), as well as the calmodulin interaction domain (^) and the transmembrane domain

Full size image

Lethal-wasting mice lack VAMP1 protein

The G190T nonsense mutation is predicted to truncate nearly half of the protein. Clustal analysis shows that the VAMP1 protein is highly conserved between species and within the protein family. VAMP1 has several important domains including a transmembrane domain and a calmodulin interaction domain that would be eliminated in the lethal-wasting mutant (Fig. 2b). However, the sharp decrease in mutant mRNA levels suggests that nonsense-mediated degradation of the VAMP1 message may occur. Western blot analysis was used to detect the presence or absence of a truncated protein. Protein samples from wild-type and mutant brain tissues were probed with an anti-VAMP1 antibody. This antibody was raised to the N-terminal 18 amino acids of the rat VAMP1, thereby allowing specificity within the family and detection of the putative lethal-wasting truncated protein. Figure 3a shows that no VAMP1 protein is detectable in the mutant brain. β-Actin was used as a loading control. The lack of VAMP1 in the lethal-wasting mutant was confirmed by immunohistochemistry.

Fig. 3
figure 3

Western blot analysis shows no detectable protein in the mutant mouse. Protein was isolated from the brain tissue of the mutant and control littermates and used for Western blot analysis (panel a). Lane 4 shows the wild-type protein at the expected size of 17 kDa (ladder, lane 1). Lane 5 shows a lack of protein in the mutant. β-Actin is shown as a loading control (lanes 2 and 3). Panel b, in agreement with the lethal-wasting phenotype, Vamp1 mRNA is expressed most highly in neuronal tissues; β-actin is shown as a control. RT–PCR was performed on DNase I treated RNA from the indicated tissues: Br brain, Cb cerebellum, Ey eye, Ki kidney, Li liver, Lu lung, Mu muscle, SI small intestine, Sp spleen, Te testis, and minus sign negative control

Full size image

VAMP1 is expressed in the retina and in discrete neuroanatomical regions

In agreement with the lethal-wasting phenotype, RT–PCR using RNA from multiple tissues shows that VAMP1 is predominantly expressed in neuronal tissues (Fig. 3b). The primers used for the RT–PCR were designed from an area of the 3′ untranslated region that is common to known alternative splice forms. The lethal-wasting mutant phenotype does not indicate which areas of the brain may be affected; therefore, we used immunohistochemistry to identify where VAMP1 is expressed. Mutant and control (P14) tissue samples were sectioned and immunostained with anti-VAMP1 antibody. In wild-type tissue, high VAMP1 expression was detected in the zona incerta and rostral periolivary region (Fig. 4), whereas other neuroanatomical regions showed negligible expression of VAMP1. In the retina, staining was observed in the outer segments of the photoreceptors, in the outer and inner plexiform layers, and in a subset of ganglion cells. Staining was punctate, excluded from the nucleus and associated with the plasma membrane. As with the Western blot, no immunoreactivity was observed in mutant tissues.

Fig. 4
figure 4

Immunohistochemistry shows that VAMP1 (green) is expressed in the zona incerta, rostral periolivary region, and retina of wild-type mice and lacking in the mutant; however, no gross histological defects are observed in the mutant mouse. The zona incerta (panel a), rostral periolivary region (panel c), and retina (panel e) of a wild-type mouse show strong immunoreactivity with the anti-VAMP1 antibody (arrows indicate ganglion cells). Panels b, d, and f demonstrate the lack of immunoreactivity in the matched sections from mutant tissues. The inset shows the association of VAMP1 with the plasma membrane. The slides were DAPI counterstained (blue). The following abbreviations are used: zi zona incerta, ic internal capsule, rpo rostral periolivary region, Il infralimbic cortex, mcp middle cerebellar peduncle, OS outer segments, ONL outer nuclear layer, OPL outer plexiform layer, INL inner nuclear layer, IPL inner plexiform layer, and GCL ganglion cell layer

Full size image

Evaluation of the differential expression of SNARE genes in the lethal-wasting mutant

Lethal-wasting mutant brains were evaluated for the misregulation of the gene expression of other neuronally expressed members of the SNARE complex (syntaxin 1a, SNAP25, and syntaxin binding protein 2) and potentially compensating VAMPs (synaptobrevin-like 1, also known as Ti-VAMP or VAMP7 and VAMP2). Quantitative RT–PCR was performed on RNA isolated from microdissected brain tissues from four mutant and four wild-type littermates. The relative expression levels (transcripts per 1,000 transcripts of β-actin) of the genes were calculated. Although all genes had slightly elevated expression levels in the mutant compared to wild-type, the differences were not statistically significant (Fig. 5).

Fig. 5
figure 5

Expression levels of other SNARE proteins are unchanged in the lethal-wasting mutant. Relative expression levels (number of messages per 1,000 messages of β-actin) are shown for each gene. Statistical significance (p value) of the difference in expression is shown below the gene

Full size image

Discussion

In this study, we report the identification of a mutation in VAMP1 that is associated with prewean lethality and neurological defects in the lethal-wasting mouse mutant. A nonsense mutation in the VAMP1 gene was identified. Several lines of evidence support the assertion that this mutation is causal. First, the drastic nature of the mutation eliminates nearly half of the protein, including regions putatively cleaved by tetanus and botulinum toxins (peptide cleavage studies were originally performed on rat proteins, which showed that VAMP1 was sensitive to botulinum toxins D, F, and G but not tetanus toxin due to a Q76V amino acid substitution; the mouse does not have this substitution). Additionally, there is a sharp reduction in VAMP1 mRNA in mutant mouse tissues compared to wild type. Further, the mutant lacks detectable protein in the brain and retina. This was observed using an N-terminal antibody that should have detected a truncated protein if it existed. The mutation that we identified inserts a stop codon in the middle of the message and would disrupt all known alternative splice forms of the gene. Importantly, the mutation was not found in the strain of origin or in a strain survey. Finally, nearly half of the genes in the disease interval had already been associated with a phenotype in the mouse. None of these were similar enough to the lethal-wasting phenotype to warrant an allelism test.

The use of qRT–PCR as a first-pass screen of candidate genes allows for the simultaneous screening of several genes. One caveat is that mutant tissue may have several genes misregulated as a result of—and not the cause of—the phenotype. In fact, in the rat, VAMP1 is downregulated in the facial motor nucleus after axotomy of the facial nerve [32]. In schizophrenic patients, levels of VAMP1 in the superior temporal gyrus were found to be significantly lower than those of unaffected controls [33]. The results of these studies indicate that abnormalities in the expression of VAMPs may be involved in several dissimilar neurological disorders. A similar phenomenon is seen with the phosphatidylinositol metabolism pathway, which is also essential for the maintenance of normal synaptic vesicle dynamics. Null mutations in genes of this pathway result in dramatic phenotypes [34, 35], but minor disruptions of the pathway, including relatively small expression differences, are thought to be associated with several neurogenetic disorders including autism [36, 37], psychiatric disorders [38, 39], and cognitive defects [40].

The lethal-wasting phenotype is relatively mild compared to that of the VAMP2 knockout mutant. The expression pattern of VAMP1 in the brain is far more restricted than VAMP2. VAMP2 is expressed throughout the mouse brain and at extremely high levels in the hippocampus and cortex, whereas relatively lower expression levels are observed in the mes- and diencephalon (GENSAT no. 47593–95). The VAMP2 knockout mouse dies at birth, whereas the lethal-wasting mutant survives 2 weeks. The reason for this phenotypic difference is not clear, but it is reflective of the fact that VAMP2 is the predominant VAMP isoform in the brain, and that VAMP1 expression is restricted to a few regions of the diencephalon and midbrain. The lethal-wasting phenotype also indicates that these neuroanatomical regions are not immediately required for life and are likely still developing during the second postnatal week.

The VAMPs are important for membrane fusion events. Although the lipid bilayer fusion events are not well understood, the protein/protein interactions that are necessary for fusion have been thoroughly studied. The VAMPs are members of the SNARE family of proteins that are directly involved in initiating membrane fusion; however, they do not appear to represent a minimal fusion apparatus. Associated proteins from the SM family (Sec1/Munc18-like) are necessary for fusion. A knockout mutation of Munc-18a leads to a complete loss of neurotransmitter release [41], whereas deletion of VAMP2 causes a dramatic reduction, but not complete loss, of vesicle fusion. SM proteins appear to act as regulators of the SNARE assembly, whereas other proteins are involved in the regulation of synaptic vesicle exocytosis. A knockout of the Ca2+-binding protein synaptotagmin in mice shows loss of synchronous release [42]. Ca2+ induces the binding of synaptotagmin to SNARE complexes, syntaxin-SNAP25 heterodimers, and both proteins individually [43, 44], as well as phospholipids [45], thus suggesting that synaptotagmin is the major coordinator of synchronous release [46]. VAMP1/2, as members of the SNARE complex, are involved in Ca2+-stimulated synaptic vesicle release; therefore, it is likely that the lethal-wasting mutant has a defect of neurotransmission in a subset of neurons.

In the human, alternative splice forms of VAMP1 have been described with different C-terminals that are involved in the differential targeting of vesicles to subcellular structures [47]. VAMP1A is expressed in the brain and targeted to the plasma membrane, whereas VAMP1B is targeted to the mitochondria and expressed in several cell lines [48]. In the mouse, an expressed sequence tag (EST) evidence exists for the presence of two splice forms equivalent to human VAMP1A and VAMP1B. The lethal-wasting mutation would affect both of these splice forms; therefore, the significance of either splice forms to the lethal-wasting phenotype would be difficult to discern without making a knockout specifically targeting the VAMP1b isoform.

VAMP1 is expressed in the retina, zona incerta, and rostral periolivary neurons. In the retina, staining was observed in the inner and outer plexiform layers, and in a subset of ganglion cells, consistent with a previous report [49]. However, we also observed staining in the outer segments. The function of VAMP1 in the retina is unclear, as the expression pattern of VAMP1 in the retina appears to be somewhat duplicated by VAMP2. It is yet unknown whether lethal-wasting mutants have visual defects; however, unlike the VAMP2 knockout mutant, electroretinogram analysis will be possible on lethal-wasting mutants. The function of the zona incerta is not completely understood. It receives input from the precentral gyrus of the cerebral cortex and cerebellum, suggesting a role in movement, which would be consistent with the lethal-wasting phenotype. The rostral periolivary region receives input from the inferior colliculus, a major auditory processing center [50]. Thus, the lethal-wasting mouse will not only be a valuable resource for the study of vesicle fusion but for the study of discrete neuroanatomical regions for which functional information is limited.