Huntington disease (HD) is caused by expansion of a CAG trinucleotide repeat in exon 1 of a novel gene. The HD protein (huntingtin) plays a critical role in early embryonic development since homozygous targeted disruption of the murine HD gene results in embryonic lethality by day 7.5. To rescue this phenotype by transgene based huntingtin expression it is therefore essential to express the protein early enough in development in the appropriate cells. Since YAC based transgenes are known to be regulated in an appropriate temporal and tissue-specific manner, we sought to rescue the embryonic lethality by breeding YAC transgenic mice expressing human huntingtin with mice heterozygous for the targeted disruption. We generated viable offspring homozygous for the disrupted murine HD gene but expressing human huntingtin derived from the YAC. This result clearly shows that YAC transgene based expression of huntingtin occurs prior to 7.5 days gestation. Additionally, we show that human huntingtin expression in YAC transgenic mice follows an identical tissue distribution and subcellular localisation pattern as that of the murine endogenous protein and that expression levels of 2–3 times endogenous can be achieved. This shows that human huntingtin under the influence of its native promoter, despite differences to the murine protein, is functional in a murine background and can compensate for loss of the murine protein. These results show that YAC transgenic approaches are a particularly promising route to producing an animal model for disorders associated with CAG expansion.
Huntington disease (HD) is an autosomal dominant progressive neurodegenerative disorder characterised by choreic movements, psychiatric disturbances and intellectual decline (1), and is associated with neuronal loss particularly evident in the caudate and putamen with selective loss of medium spiny neurons (2).
The HD gene (IT15) was recently cloned and was found to contain a highly polymorphic CAG trinucleotide repeat in exon 1 (3) which when expanded beyond 35 repeats is associated with the HD phenotype (4). cDNA based transgenic mice expressing high levels of mutant human HD mRNA but not protein do not develop an HD phenotype suggesting that neurotoxicity is not mediated at the DNA or RNA levels but rather necessitates translation of the protein (5). The finding of embryonic lethality in mice with targeted disruption of both alleles (6–8) together with the existence of adult humans homozygous for CAG expansion in the HD gene, suggests that HD is caused by a gain-of-function at the protein level and is not due to simple haploinsufficiency.
The HD gene is highly conserved during evolution and has been identified with a high degree of homology in organisms as distantly related to humans as the puffer fish (9). We have previously cloned and sequenced the murine homologue (HDh) and shown that the two genes are 90% identical at the amino acid level (10). Comparative analysis of the 5′ genomic sequences also shows a high degree of nucleotide identity (79%) between the first 200 bp of the promoter (11). One noticeable difference between the two genes, however, is the CAG trinucleotide repeat size which is 7 in the mouse, while mean length observed on normal human chromosomes is 18. In addition, the mouse CAG is interrupted at the third triplet by a CAA (11).
An animal model for HD would greatly facilitate our understanding of the cellular and molecular mechanisms of pathogenesis. It is likely that transgene based expression of the mutant human protein in mice could result in selective neuronal loss depending on the promoter used but the limitation to this approach is that the expressed protein is likely to be developmentally regulated in a tissue-specific fashion dependent on the promoter. For example, the development of an ataxic phenotype in mice expressing the mutant SCA-1 protein from a cDNA under control of a Purkinje cell specific promotor, PCP-1, represents an example of polyglutamine induced toxicity in these cells, following levels of expression significantly greater than endogenous levels (12). Furthermore, a similar phenotype is seen when the expanded polyglutamine tract of the gene causing SCA-3 (Machado-Joseph disease) is expressed under control of the same PCP-1 promoter (13). In this instance, only a few residues besides the polyglutamine tract were included in the construct suggesting that the surrounding gene sequence does not play a significant role in mediating the neurotoxicity.
The development of transgenic technologies using YACs offers important advantages for generating transgenic mice expressing human huntingtin. The inclusion of endogenous regulatory elements within the transgene should allow for the normal temporal and tissue specific expression of human huntingtin (14–16). Furthermore, expression of transgenes from YACs reach levels comparable with that of the endogenous gene in a copy number dependent and position independent manner (17,18). Often, cDNA based transgenes are poorly expressed and inappropriately regulated (19). In addition, site directed mutations can be introduced into YACs using a homologous recombination based strategy in yeast (20). This is extremely important since introducing expanded CAG repeats into the transgenes will be essential for the development of neuropathology.
To test whether YACs would be appropriate vectors for expression of the human HD gene in mice, we sought to generate YAC transgenic lines containing the normal HD gene including all endogenous regulatory sequences. We have generated a number of transgenic lines containing intact YACs and demonstrate that the human huntingtin is expressed at levels and is developmentally regulated in a tissue-specific manner that clearly parallels that of the mouse endogenous protein.
Additionally, we investigated whether human huntingtin would be functional in the murine background and that the murine genes necessary for appropriate post-translational modulation and transport of murine huntingtin would be equally effective with the human protein.
In an effort to assess whether human huntingtin expressed by YAC transgenic mice can functionally substitute for the murine HD protein, we bred YAC transgenic mice expressing human huntingtin with mice containing a targeted disruption of the murine gene to generate offspring homozygous for this disruption, but containing the transgene. Interestingly, the mice with homozygous gene disruption which normally are not liveborn, are rescued by human huntingtin derived from the YAC, conclusively proving that these YACs are developmentally expressed early, prior to day 7.5 of gestation similar to the murine protein, and that human huntingtin can compensate for loss of the murine protein. These results suggest that the YAC transgenic approach offers an important potential route to producing an animal model representative of human HD.
Generation and identification of HD YAC transgenic mice
To generate transgenic mice expressing the human HD gene, we microinjected two previously characterised YACs, YGA2 and 353G6 (21). Both contain the entire HD gene with 18 CAG repeats and show no obvious rearrangements based on extensive mapping analysis (21) but differ with respect to size and flanking DNA sequences (Fig. 1). YGA2 is 600 kb in size and extends ∼350 kb 5′ and 50 kb 3′ of the gene whereas 353G6 is 350 kb, extending 25 kb 5′ and 120 kb 3′ of the gene.
Purified YACs were microinjected into 871 (YGA2) and 346 (353G6) pronuclei derived from the FVB/N mouse strain and transplanted into foster mothers (see Materials and Methods). To assess YAC integrity and to identify positive transgenic founder mice, we screened mouse tail DNA with five PCR markers that span the entire length of the YACs for evidence of successful YAC integration (Fig. 1). Primer pairs from the left (LYA) and right (RYA) YAC arms (22) and intragenic primer pairs from exon 1 (CAG), intron 1 (CA repeat), and exon 59 (ΔG) were used. A total of 8/54 (15%) potential YGA2 founders contained YAC sequences whereas 11/28 (36%) of potential 353G6 founders showed evidence of YAC integration (Table 1 and Table 2). Four positive YGA2 founders (mice # 40, 44, 45 and 46) contain fragmented YACs with only 5′ HD gene sequences. All but two of the positive 353G6 founders integrated the YAC intact (Table 2). All founders transmitted the YAC (except founders 1 and D207 who failed to breed) with a frequency of transmission usually between 20–30% suggesting mosaicism of the germline (22). Breedings of F1 progeny, however, resulted in the expected 50% frequency of transmission.
Assessment of YAC integrity and copy number
To further assess the integrity of the YAC in positive transgenic founders, Southern blotting and fluorescence in situ hybridisation (FISH) were used. Southern analysis was done on EcoRI digested mouse and human genomic DNA probed with a 4.1 kb EcoRI fragment encompassing exon 1 and the regulatory region of the HD gene. All lines tested (Table 1 and Table 2) showed a 4.1 kb band revealing that the regulatory region and exon 1 of the human gene within the YAC was intact (data not shown). Copy number was estimated by comparing the hybridisation intensity of the 4.1 kb band in each line with human genomic DNA. Most lines integrated a low number of YAC copies (1,2) with the exception of line B60 which had integrated ∼2–3 copies and D212 (4–5 copies).
Chromosomal location of the HD transgene
The HD transgene was mapped to metaphase chromosomes of most transgenic lines by FISH using a 35 kb human cosmid probe from the 5′ end of the gene (L83d3) (21). The results are summarized in Table 1 and Table 2 and examples of hybridization are shown in Figure 2. Between 15 and 63 cells were scored for each line. A double hybridisation signal was observed on each chromatid for mouse lines D212 and B60 confirming a higher copy number. No other significant hybridisation signals were noted in each line, indicating that the YACs had integrated into one site in the mouse genome.
RNA expression analysis
Since our primary goal was to create transgenic mice expressing the human HD gene, we checked for human mRNA expression using reverse transcription polymerase chain reaction (RT-PCR). RNA was extracted from mouse tails and following the initial RT reaction, human specific PCR primers from exon 7 and exon 9 were used to amplify human transcript derived from the YAC. Primers corresponding to sequences in exons 64 and 66 were also used to check for expression of the full length mRNA. All founders that had tested positive for the PCR screens were shown to express the full length human HD transcript (Fig. 3) with the exception of the YGA2 founders containing only the 5′ end of the HD gene. These mice express a truncated transcript detected only by the exon 7 and exon 9 primer pair sets. 353G6 founder D211 shows no evidence of expression but may express a shorter transcript that would not be detected by the primers used here.
Human huntingtin expression analysis
For all protein studies performed on the YAC transgenic mice, we used two primary antibodies: (i) GHM1, a monoclonal antibody generated from amino acid residues 1678–1981 that detects human but not murine huntingtin on western blots and (ii) a polyclonal amino terminal antibody, BKP1, generated from amino acid residues 3–16 which detects both mouse and human huntingtin on western blots with equal proficiency (23). The 350 kDa band detected by both antibodies was shown to be competed out upon incubation with the appropriate antigen.
To detect the presence of human huntingtin in the transgenic lines, protein from cortical tissue of the F1 transgenic mice was analysed by western blotting using GHM1. Human cortex and human 293 kidney cell lysate, and non-transgenic mouse cortical tissue were used as positive and negative controls respectively. Human huntingtin expression was detected in all lines known to be expressing human mRNA (Fig. 4a) with line D212 showing the highest levels of expression.
To compare the levels of expression between the human huntingtin in line D212 and endogenous murine HD, equal amounts of protein were probed on western blots with BKP1 as the primary antibody and densitometry scanning was performed (Fig. 4b). Human huntingtin expression levels in line D212 were estimated to be 2–3 times that of mouse endogenous protein. The high level of transgene expression in this line can be attributed to the increased transgene copy number which we estimate to be 5–6 copies. The same analysis was performed for line 30 showing that transgene protein expression levels were equivalent to murine endogenous levels for this line (data not shown).
Tissue distribution of YAC and murine derived huntingtin
It has been shown that the pattern of gene expression from human YACs in a mouse background closely mimics that of the corresponding endogenous murine gene for genes such as the amyloid precursor protein (18) and the β-globin genes (14). To test whether this was true for the YGA2 and 353G6 YACs, human and murine derived huntingtin expression levels were assessed in brain and peripheral tissues which included cortex, cerebellum, heart, lung, liver, spleen, kidney, testis and skeletal muscle. Equal amounts of total protein from each tissue were compared on western blots using GHM1 for transgenic line D212 (YAC 353G6) and BKP1 for normal mouse tissues.
As seen in Figure 5, tissue expression patterns of human and mouse huntingtin are identical with the highest expression levels detected in the cortex, cerebellum and testes. A longer exposure showed that huntingtin was expressed in all other peripheral tissues tested, albeit at much lower levels (data not shown). An immunoreactive band of ∼150 kDa in size was detected in skeletal muscle when probed with GHM1 but did not appear when probed with BKP1. This band could be eliminated in competition tests using excess antigen (data not shown). This indicates that the observed band could be a C-terminal cleavage product of the HD gene produced primarily (if not exclusively) within skeletal muscle. An identical tissue distribution pattern was observed for YGA2 line 30 (data not shown).
To assess the subcellular localisation of huntingtin in the YAC transgenic mice, cortical tissue from a normal and a transgenic mouse (line D212) was fractionated into P1 (nuclear), P2 (mitochondrial), P2S (synaptosomal), P3 (microsomal and plasma membranes) and S3 (soluble) fractions and subjected to western blotting (Fig. 6). The results indicate that mouse and YAC derived human huntingtin are predominantly localised to the soluble cytosolic fraction with much lower detected levels of immunoreactivity in the P3 and P2S fractions. Similar results were obtained using cerebellum and testis. However, in the testis, increased levels of murine and human huntingtin were detected in the P3 fraction. Western blot analysis was done on the same fractions from mouse cortex using a monoclonal antibody that recognizes synaptobrevin, an integral membrane protein of synaptic vesicles (24). Synaptobrevin was highly enriched in the P2S fraction with limited or no immunoreactivity detected in the S3 or P3 fractions, respectively (data not shown).
The localisation of huntingtin was compared in multiple brain regions from both normal and YAC transgenic mice. Frozen sections including frontal cortex, caudate/putamen, globus pallidus, parietal cortex, hippocampus, thalamus, cerebellum, brainstem and spinal cord were examined using BKP1. In each of these regions, immunostaining with BKP1 revealed cytosolic immunoreactivity strongest in neuronal cells, with a lower intensity of staining seen in the cytosol of glial cells (Fig. 7A–C). No immunoreactivity was seen in the nuclei. The specificity of the demonstrated immunostaining was confirmed by the ineffectiveness of this antibody on sections from both normal (Fig. 7D) and YAC transgenic (data not shown) mice after preincubation with excess BKP1 antigen.
The distribution of immunostaining was identical between normal and YAC transgenic mice. However more intense staining was observed in the YAC transgenics, indicative of expression of human huntingtin. This data clearly shows that the human and mouse proteins are localised similarly in the cytosol, predominantly in neurons, in all brain regions observed.
Identical results were obtained with the human specific antibody MAB 2168 in the transgenic mice (data not shown). No or little immunoreactivity was observed in normal mice, further indicating that the YAC transgenic mice are expressing human huntingtin.
Postnatal developmental expression
To determine whether human huntingtin expression from the YAC was maintained and regulated during postnatal development similar to normal mice, we measured protein expression in transgenic mouse cortex (D212) at 3, 6 and 12 weeks of age by western blotting using GHM1 (Fig. 8). Human HD protein was detected at each stage showing no obvious changes in levels of expression and with patterns of expression similar to that seen for mouse protein (data not shown).
Rescue of the HD knock-out lethal phenotype
Since our data showed that the YAC was regulated in a manner that was the same as that of the mouse endogenous gene, we hypothesized that human huntingtin would also be expressed at an early stage of development in the appropriate cells and that this human protein could rescue the embryonic lethal phenotype observed in mice homozygous for disruption in the murine HD gene (6–8).
YAC transgenic mice from lines 29 and 30 (YGA2) and lines B50 and B60 (353G6) were bred to mice heterozygous for the targeted disruption in the mouse HD gene (6). A total of 67 of the offspring tested positive for the YAC and were genotyped at the mouse HD locus using a mouse specific probe derived from intron 5 (6) (Fig. 9; Table 3). All four lines demonstrated efficient rescue of the embryonic lethal phenotype including line 29 which we estimate to be expressing huntingtin levels approximately half that of mouse endogenous huntingtin. This clearly shows that low levels of huntingtin are sufficient to induce rescue of embryonic lethality. Of the remaining 52 non-transgenic offspring, none were homozygous for the targeted disruption (data not shown). This rescue implies that human huntingtin was appropriately expressed in the correct cells prior to day 7.5 of gestation from both YACs. Furthermore, the human protein functions similarly to the mouse protein in development.
Behaviour of five YAC transgenic mice which were also homozygous for the targeted disruption in the HD gene appeared to be no different from their littermates at four months of age. They ate and drank normally and were indistinguishable from controls with respect to body weight, posture, locomotion, rearing and grooming and did not display any signs of ataxia or neurological abnormalities. However, no detailed morphometry or formal behavioral assessment of these mice has been undertaken at the present time.
We have generated 13 YAC transgenic lines, each carrying copies of YACs (or parts thereof) encompassing the HD gene which spans 210 kb of genomic sequence. YGA2 is a 600 kb YAC, one of the largest YACs to be microinjected and recovered intact whereas 353G6 is smaller in size at 350 kb. Line D212 (derived from YAC 353G6) which demonstrated a double hybridisation signal by FISH and a much greater signal intensity than human genomic DNA by Southern blotting indicative of an increased copy number, had the highest levels of expression which was estimated to be 2–3 times that of the mouse endogenous protein. Line B60 had 2–3 copies of the transgene, double hybridisation signals by FISH, and also had higher levels of huntingtin expression compared with other lines, but less than line D212. In most other transgenic lines with lower copy numbers (1–2), human protein levels were comparable with murine huntingtin levels. This data is consistent with the notion that YAC transgene expression is comparable with endogenous levels and is largely determined by copy number (17,18).
Human huntingtin expressed from the YAC transgenes demonstrated identical tissue distribution as well as cellular and subcellular localisation as endogenous mouse huntingtin. Expression patterns of human and mouse huntingtin was highest in the brain and testes with significantly lower levels in peripheral tissues. The size of the protein was similar in all tissues (∼350 kDa) besides skeletal muscle which, when probed with a human specific C-terminal antibody, revealed a smaller product of 150 kDa. This band was observed in all lines tested (D212 and 30). It is unlikely that this is a cross-reactive band as it was competed out with appropriate antigen and suggests that in muscle a smaller C terminal product of the HD gene is present. This could be the result of alternative splicing, a different translational initiation site or tissue specific proteolytic processing.
Immunohistochemical analysis revealed that human and murine huntingtin localised to the cytosol. More intense staining was observed in the YAC transgenic mice, indicative of expression of both murine and human huntingtin. In the brain, huntingtin was predominantly expressed in neurons. Furthermore, analysis of the subcellular localisation of human and murine huntingtin revealed that both mouse and YAC derived huntingtin are primarily localised to the soluble cytosolic fraction. To assess postnatal developmental expression of YAC derived human huntingtin compared with the endogenous murine protein, expression was assessed in the transgenic mouse cortex at 3, 6 and 12 weeks of age. Again, a similar pattern of expression was seen for the protein derived from the transgene compared with endogenous protein.
We have previously shown that murine huntingtin is crucial for the elaboration of mesoderm and when absent results in embryonic lethality at around day 7.5 (6). In an effort to assess whether YAC derived human huntingtin was also appropriately expressed prenatally, we bred YAC transgenic mice with mice heterozygous for the targeted disruption in the mouse HD gene. A total of 15 of 67 mice were liveborn but were also homozygous for the targeted disruption of the HD gene. Therefore, human huntingtin derived from both YAC transgenes (353G6 and YGA2) was appropriately expressed prior to day 7.5 of gestation, and furthermore could compensate for the loss of the mouse protein in development. This also suggests that the human huntingtin undergoes appropriate post-translational modification and can interact with other murine proteins which might be crucial for its normal functions. These results also show that despite a 10% difference in amino acid residues (10), this does not interfere with the ability of human huntingtin to substitute for the murine protein early in development.
In summary, these results clearly show that the YAC derived transgene expression of the human protein is appropriately expressed during prenatal and postnatal development and that the human protein is localised to similar subcellular compartments as murine huntingtin.
We have previously shown intergenerational stability of the expanded CAG repeat in transgenic mice containing the HD cDNA which suggested a potentially important role of genomic sequences in mediating trinucleotide instability (5). Here we have established YAC transgenic mice containing between 150–350 kb of 5′ and/or 3′ sequence flanking the HD gene. Analysis of ∼100 offspring has revealed that there is no evidence for CAG intergenerational instability. We have previously shown in humans that CAG intergenerational instability for CAG size of ∼18 is extremely rare (25,26). Therefore the CAG stability in the YAC transgenic mice suggests that larger CAG sizes are necessary for CAG instability. Establishment of YAC transgenic mice with larger CAG sizes will allow us to assess whether the inclusion of surrounding genomic sequences in the presence of a larger CAG size confer CAG instability in this gene.
The findings presented in this paper have important implications for the development of an animal model for HD. To reproduce the phenotype of HD in mice, the gene containing a CAG repeat size >36 should be appropriately expressed during development in the correct tissues and localised in subcellular compartments in a manner similar to the endogenous protein. cDNA transgenic approaches may provide high levels of protein expression in specific tissues, but are unlikely to truly reflect the HD phenotype as they are unlikely to be correctly regulated. Therefore YAC transgenic approaches provide an important alternative route to the production of an animal model for HD. Furthermore, YACs can be modified using homologous recombination in yeast to introduce large CAG repeat lengths (20). In such an instance the expanded CAG remains under the control of the native promoter and the more distal regulatory regions are maintained.
Using transgenic approaches and YAC DNA, we have created a mouse that over-expresses the human huntingtin gene product. Our results show that YAC transgenic approaches may be a particularly fruitful route to produce an animal model representative of human HD. In addition, as it is postulated that HD results from a gain-of-function, these mice, even at this stage, could represent an important tool to test therapeutic approaches aimed at reducing the level of expression of human huntingtin.
Materials and Methods
YAC purification for microinjection
YAC DNA was prepared for microinjection essentially as described (17) with a few modifications (22). High density yeast plugs were prepared by growing yeast cultures to saturation at 30°C in a 500 ml culture of ura- trp- dropout media. After washing, cells were resuspended in 4 ml SCE (1 M sorbitol, 0.1 M Na citrate, 5 mM EDTA) mixed with 6 ml of 3% InCert agarose (FMC) at 42°C and dispensed into plug molds (Biorad). Once solidified, plugs were incubated overnight at room temperature in 10 ml SCE with 10 mg of lyticase (Sigma). This solution was replaced the next day with ES (0.5 M EDTA, 1% sarkosyl) with proteinase K (1 mg/ml) and placed at 50°C overnight. Plugs were then washed five times with T10E50 (10 mM Tris pH 8.0, 50 mM EDTA) for 30 min incubations and stored at 4°C. High molecular weight DNA was prepared for microinjection as described (22) except that the running conditions of the pulsed field gel electrophoresis (PFGE) were modified for YGA2 and 353G6. YGA2 was separated at a 120° reorientation angle at 6.0 V/cm with a 60 s switch time for 15 h followed by a 120 s switch time for an additional 12 h whereas 353G6 was separated at a 120° reorientation angle at 4.0 V/cm with a 50 s switch time for 30 h.
Analysis of transgenic founders
DNA was extracted from mouse tails for PCR analysis as described (27). Primer pairs were used to identify YAC sequences from the left and right YAC arms (22), the CAG repeat exon 1 (28), the CA repeat in intron 1 (this paper), and the ΔG in exon 59 of the HD gene (29) using conditions as described. The CA repeat was amplified using 800 mM of primer IN1-1 (5′-TGA GGC TGC AGT GAG CTA TGA-3′) and 800 mM (200 mM α-32P end-labelled and 600 mM cold) of primer In1–2 (5′-CTC ACC AGC ATG TGG TAT TG-3′) in a 25 µl volume containing 2 mM MgCl2, 100 mM each dNTP, 1× PCR buffer (BRL), and 7% DMSO. DNA was denatured for 1 min at 95°C, followed by 7 cycles of touchdown PCR: 94.5°C for 40 s, annealing decreasing by 1°C per cycle from 63°C for 40 s, followed by extension at 72°C for 50 s. This was followed by 30 cycles of 94.5°C for 40 s, 59°C for 40 s, and 72°C for 50 s and a final extension of 72°C for 10 min. PCR products were sized against an M13 sequencing ladder on a 6% urea-polyacrylamide gel and subjected to autoradiography.
A total of 10 µg of transgenic mouse genomic DNA and human genomic DNA was digested with EcoRI (BRL), run at 50 V for 14 h in a 1% TBE gel and blotted to Hybond-N+ membrane (Amersham). Fifty nanograms of the 4.1 kb EcoRI fragment derived from the 5′ region of the HD gene (encompassing exon 1 and the regulatory elements) was radiolabelled with [α-32P]dCTP using random primers and used as a probe. Sheared human placental (500 µg) and salmon sperm (2 mg) DNA were added to the hybridisation mix to prevent non-specific binding. Filters were washed at 60°C for 20 min in 0.5× SSPE, 0.1% SDS, and an additional 20 min in 0.1× SSPE, 0.1% SDS and finally exposed to Kodak X-OMAT AR film.
RNA was extracted from mouse tails using the guanidium isothyocyanate method (30). Random hexamer primed RNA (5 µg) was reverse transcribed using the Superscript preamplification system following manufacturers instructions (BRL). Aliquots of the reaction mix (1 µl) were PCR amplified using 800 mM each of human specific primers HD1201 (5′-GTG CTC TTA GGC TTA CTC GTT CCT TG-3′) from exon 7 and HD1481 (5′-GGC GGT CTG AAG AGC TGC TGC AAC-3′) from exons 8 and 9 in a 25 µl volume with 1× PCR buffer (BRL), 2 mM MgCl2, 200 mM each dNTP and 0.5 U Taq polymerase. A two-step PCR was performed following a 94°C denaturation step for 1 min with 35 cycles at 94°C and 72°C, both for 30 s, and a final extension time of 10 min at 72°C. Products were analysed on a 1% TBE Ultrapure agarose gel (BRL).
Fluorescence in situ hybridisation
A cosmid probe with a 35 kb insert containing the human HD gene was labeled with biotin-11 -dATP by nick translation (BRL). The size of the product was determined to be between 200 and 400 bp. Metaphase chromosome preparations from nine transgenic FVB mice were obtained using 0.075 M KCl as a hypotonic buffer and methanol:acetic acid (3:1, v/v) as fixative. The hybridisation was carried out as previously described by Chance et al. (31). Hybridisation signals were detected using a detection system from Vector. After blocking with goat serum and an incubation with fluorescein avidin DCS, slides were rinsed in modified 4× SSC (4× SSC, 0.03% Triton). A second incubation with biotinylated anti-avidin D and a rinse in modified 4× SSC followed. A final incubation with fluorescein avidin DCS and a rinse in modified 4× SSC completed the detection. The chromosomes were banded using Hoechst 33258-actinomycin D staining and counterstained with propidium iodide. The chromosomes and hybridisation signals were visualized by fluorescence microscopy, using a dual band by-pass filter (Omega).
Brain tissues were obtained from two YAC transgenic mice (lines B60 and D212) and two normal control mice, which were perfusion-fixed with 4% v/v paraformaldehyde/0.01 M phosphate buffer (4% PFA) under deep anesthesia with halothane. The brain tissues including spinal cord were removed, immersion-fixed in 4% PFA for 1 day, washed in 0.01 M phosphate buffered saline (PBS) for 2 days, and then equilibrated in 30% w/v sucrose/0.01 M phosphate buffer for 2 days. The samples were then snap-frozen in Tissue Tek molds by liquid nitrogen. After warming to −20°C, frozen blocks derived from frontal cortex, caudate/putamen, globus pallidus, parietal cortex, hippocampus, thalamus, cerebellum, brainstem and spinal cord were cut into 20 µm sections for immunohistochemistry.
Following soaking in 10 mM PBS (pH 7.2), the tissue sections were blocked using 2.5% v/v normal goat serum for 60 min to eliminate non-specific binding. Primary antibodies diluted with PBS (pH 7.2) were applied to the sections overnight at 4°C. Optimal dilutions were determined to be of 1:50 for BKP1 and 1:500 for MAB 2168 (Chemicon). This incubation was followed by sequential incubation of samples with biotinylated secondary antibody and then an avidin-biotin complex reagent (Vectastain Elite ABC Kit: Vector) at room temperature for 45 min each and the color was developed using 3–3′-diaminobenzidine tetrahydrochloride and ammonium nickel sulphate.
To confirm the specificity of the observed immunoreactivity, sections were treated as mentioned above except that the primary antibodies were replaced with normal rabbit serum or normal mouse serum. As an additional negative control, BKP1 was preabsorbed with excess peptide prior to incubation with the tissue sections.
Subcellular fractionation of brain tissue
Cortical matter from frozen mouse brains was fractionated. Cortical matter was separated from brainstem, cerebellar and nuclear regions of whole brains, immediately frozen in isopentane, cooled in liquid nitrogen and stored frozen at −70°C. Brain tissue was thawed and homogenized on ice in 10 vol sucrose buffer (0.303 M sucrose, 20 mM Tris-HCl pH 7.2, 0.5 mM EDTA, 1 mM MgCl2, 1 mM PMSF, 1 mM benzamidine, 5 µ/ml leupeptin and 10 µg/ml soybean trypsin inhibitor) using a glass-teflon IKA-RW 15 homogenizer (Tekmar Company, Cincinnati, Ohio) at maximum speed. The homogenate was fractionated by a differential centrifugation protocol at 4°C as described (32). Centrifugation of the homogenate for 5 min at 1300 g gave rise to the nuclear and cellular debris pellet (P1). The supernatant (S1) was saved and pooled with a second S1 fraction derived from re-homogenizing the P1 pellet. The S1 supernatants were centrifuged for 20 min at 11 750 g to give a mitochondrial/synaptosomal P2 pellet. The cream-coloured upper portion of P2 (P2S), enriched in synapto-somes, was crudely removed from the lower brown-coloured portion containing mitochondria (P2). The S2 supernatants were centrifuged at 142 000 g for 35 min to pellet micosomal (ER and Golgi) and plasma membranes (P3). The final supernatant from this step (S3) contained cytosolic protein. The P1, P2 and P2S membrane fractions were washed twice in sucrose buffer and then solubilized by heating for 20 min in 3 M urea, 1% SDS, 1 mM DTT, TBS, pH 8. Protein was determined by a modified Lowry protein assay (Bio-Rad DC Protein Assay, Bio-Rad) and 80 µg of each fraction was analysed by western blotting.
Tissue distribution of huntingtin
Frozen tissue samples were homogenized in sucrose homogenization buffer (see protocol above). Brain tissue was homogenized with a glass-teflon homogenizer and peripheral tissues were homogenized using a Polytron. Homogenates were sonicated for 3 × 5 s at maximum percentage output on a Microson utrasonic cell disrupter (Heat Systems Ultrasonics, Farmingdale, NY) and then centrfuged for 5 min at 1300 g to precipitate cellular debris. Total protein was assayed and 200 µg of each tissue fraction was mixed 4:1 in sample loading buffer and analysed by western blotting.
Proteins were diluted 4:1 in sample loading buffer (250 mM Tris-HCl pH 6.8, 10% SDS, 25% glycerol, 0.02% bromophenol blue and 7% β-mercaptoethanol) and resolved by SDS-PAGE using 7.5% mini gels (Bio-Rad Mini-PROTEAN II Cell system) as described (33). For the analysis, huntingtin proteins were electroblotted overnight in transfer buffer containing 10% methanol at 30 V onto PVDF membranes (Immobilon-P, Millipore) (34). Anti-synaptobrevin antibodies were kindly provided by Dr Reinhard Jahn, Howard Hughes Medical Institute.
Rescue of embryonic lethality
FVB/N mice expressing the YAC transgene were mated to mice heterozygous for the murine HD gene disruption. Tail DNA was extracted from offspring and screened using the RYA PCR (22) and genotyped at the murine HD locus for the null allele using primers P8 and P9 (6). 1× PCR buffer (GIBCO BRL), 2 mM MgCl2, 800 mM of each primer and 0.5 U Taq polymerase were mixed in a 25µl reaction and subjected to PCR: denaturing at 95°C for 1 min, 35 cycles at 94°C, 62°C and 72°C for 60, 45 and 45 s, respectively, followed by 10 min extension at 72°C. Heterozygous mice containing the YAC were bred to heterozygous littermates. Offspring from this second breeding were screened again with the RYA PCR and positive mice were genotyped by Southern analysis as described (6) using an α-32P radiolabelled 540 bp Xba/HindIII fragment from mouse intron 5 as a probe on EcoRI digested mouse tail DNA. Filters were exposed to Kodak X-OMAT overnight at −70°C.
We thank G. Bates for the gift of the YACs. This work was supported by grants from the MRC (Canada), the Canadian Genetic Diseases Network, the March of Dimes Birth Defects Foundation (to C.D.) and the National Institutes of Health (to C.D. and M.B.D.). We thank S. Thomas for her assistance and Dr Reinhard Jahn for the anti-synaptobrevin antibodies.