Insertional Mutation in the Intron 1 of Unc 5 h 3 Gene Induces Ataxic , Lean and Hyperactive Phenotype in mice

Mice carrying a mutation in the first intron of Unc5h3 were accidentally produced by transgenic insertion and characterized for their homozygous mutant phenotypes. Morphological and histological analysis revealed cerebellar and midbrain abnormalities, which are similar to the previously reported phenotypes of the Unc5h3 mutant. Behavioral analysis showed higher ambulatory activity and circling, and defects in habituation in a novel environment. Their body weights were 10–30% less than wildtype mice from 2–3 weeks of age to 22 months possibly due to reduced accumulation of adipose tissues. The transgenic insertion site was identified and mapped to the intron 1 of Unc5h3 gene with approximately 50 kb deletion of the intron sequence. This intronic mutation interfered with the mRNA expression of the Unc5h3 gene not in testes, but in many tissues including the brain, implying that this intronic region may play a role in regulating tissue-specific transcription of Unc5h3.


Introduction
The netrin ligands that bind with high affinity to UNC5 as well as to Deleted in Colorectal Cancer (DCC) family receptors, guide neuronal growth cones in the developing nervous system [18,22,23].While DCC family receptors have been implicated in both attraction and repulsion, UNC5 family receptors are required only for repulsion [14,19,29].During the development of the nervous system, netrin1 acts not only as a guidance cue but also as a survival factor via its recep-normal littermates but usually catch up in body weight by 5 or 6 weeks of age [21].Besides their mild ataxic movement, they also ambulate extensively [8].
One of the major roles of the cerebellum is to regulate the motor system, which includes balance, eye movements, body and limb movements, movement planning and motor learning.Damage to the cerebellum does not alter sensory thresholds or the strength of muscle contraction but impairs the balance and the fine control of movement [10].Cerebellar ataxia results from several genetic alterations and non-genetic lesions that cause degeneration of the cerebellum [4,7,11,17,20,27,28,30]. Ataxia also results from abnormal development of the cerebellum due to defects in neuronal migration and proliferation [1,8].
In addition, the cerebellum is involved in spatial processing and learning, habituation of exploration behavior, perception of time, and integrated autonomic functions.There is a growing body of evidence that shows the cerebellum is implicated in cognitive and emotional functions [2,5,12].Early lesions of cerebellar vermis is manifested as autistic-like symptoms such as hyperactivity, persevering behavior, absence of attention to environmental distractors and disinhibition tendencies [5].Autistic patients show cerebellar abnormalities, particularly, hypoplasia of vermal lobules VI and VII.Attention deficit hyperactivity disorder (ADHD) patients have smaller cerebella especially in the posterior vermis compared with normal controls [3,25].
A new recessive cerebellar mutation arose accidentally in the process of producing transgenic mice for Tert (telomerase reverse transcriptase), which shows phenotypes similar to those of the Unc5h3 rcm mutant mouse previously published [1].Here we further investigated the phenotypes of the new mutant, including cerebellar histology, locomotor activity and growth.We also identified the transgenic insertion which creates a deletion in intron 1 of the Unc5h3 gene and studied the effects of the mutation on the expression of the Unc5h3 gene.

Mice
We generated seven lines of transgenic mice by microinjecting the 5.6 kb transgenic construct for the expression of Tert cDNA under the control of chicken beta-actin promoter, into fertilized eggs of the inbred strain FVB/N.Ataxic mice were observed in some of the offspring from hemizygous matings of only one of the transgenic founders (line 5).The offspring were analyzed for inheritance of the transgene by the phenotypes (i.e.ataxia) and PCR genotyping.The phenotypes were inherited in a recessive mode, with no effects observed in hemizygous littermates.

Histological analysis
Adult mice were anesthetized and perfused through the heart with ice-cold PBS, followed by 4% paraformaldehyde in 0.1 M PBS.The brains were removed, post-fixed in the same fixative overnight at 4°C and cryosectioned at -20°C.Sections were dehydrated and then stained with cresyl violet (Sigma-Aldrich, USA).

Open-field locomotor activity
Mice were placed in a Plexiglas chamber (60 × 60 × 45 cm) and their locomotor activities were assessed by tracking horizontal movements for 2 h.A video tracking system, which includes an overhead camera and image analysis software, SMART (Panlab s.l., Spain) was used to monitor activity.

Food and water intake
Adult mice were singly housed in cages, provided with food and water ad libitum, and evaluated every 3 or 4 days for 2 weeks for food and water intake.Relative daily food and water intake were calculated by dividing the total weight of food and water intake with the number of days and the body weight of each mouse.

Identification of the transgene insertion site by inverse PCR
Cloning of the insertion sequence was performed with inverse PCR with some modifications.Genomic DNAs isolated from the homozygous mutant mice (line 5) showing typical phenotypes and other Tert transgenic mice of a different line (line 27) were digested with EcoRI, PstI or HinfI restriction enzymes.Digested DNAs were separated on a 1 or 1.5% agarose gel by electrophoresis and blotted onto Hybond-N+ membrane (Amersham Biosciences, USA).Southern hybridization was performed with three different probes specific to the transgenic construct and the band pattern of the mutant mice was compared with that of line 27 transgenic mice.The DNA bands that are unique to the mutant mouse were isolated from the gel and self-ligated at 16°C overnight.Inverse PCRs were performed using combinations of sense and anti-sense primer pairs that are specific to the transgenic construct, sense (pCA5F1, 5'-GGCCCTATAAAAAGCGAAGC-3'; pCA5F2, 5'-GCTCGTTTCTTTTCTGTGGC-3') and anti-sense (pCA5R1, 5'-GGAAAGTCCCTATTGGCGTT-3'; pCA5R2, 5'-TCACCTCGACCC ATGGTAAT-3'), with circularized restriction fragments (2.5 and 2.0 kb EcoRI digested fractions; 1.7 kb HinfI fractions; 6.0, 3.0 and 1.8 kb PstI fractions) as templates.One of the flanking sequences was amplified by PCR from the 1.7 kb HinfI fractions, and confirmed by Southern analysis and direct sequencing.The sequence was used to search the ENSEMBL mouse genome database (v.7.3b.3)for the identification of the genomic locus using the BLAST program at the ENSEMBL BLAST server (http:// www.ensembl.org/Mus_musculus/blastview).

RT-PCR
Total RNAs from various tissues of neonatal and adult mice were prepared with Trizol reagent (GIBCO BRL, USA).To remove genomic DNA contamination, isolated RNAs were incubated with DNaseI for 1 h at 37°C.The reverse transcription reaction was carried out using 5 µg of total RNA with Superscript II reverse transcriptase (GIBCO BRL, USA) at 42°C for 60 min after hybridization with 100 ng of 6-nt random hexamer.As a control, reactions were also carried out in the absence of reverse transcriptase.Amplification was carried out using Taq polymerase for 30 or 35 cycles, each cycle consisting of 94°C for 30 s, 58°C for 30 s, and 72°C for 30 s.A primer pair (sense, 5'-GAACATCAAGGCTGCCAGA-3' and anti-sense, 5'-GCAGAGCATAGACAGGTCCC-3') was used to amplify Unc5h3 cDNA.To normalize the amounts of input RNA, amplification of glyceraldehyde-3-phosphate dehydrogenase (Gapdh) RNA was used as control.

Generation of the transgenic mutant line
Severe ataxic movement was observed in some of the offspring from intercrosses of one (line 5) of seven Tert transgenic lines, which express Tert cDNA under the control of chicken beta-actin promoter.These homozygous mice (tg/tg) were smaller in size and hyperactive compared with wildtype (WT) and hemizygous mutant mice (tg/+).Gross morphological examination indicated protrusion of colliculus, and reduction in cerebellar size and in numbers of cerebellar lobules (Fig. 1A).Histological analysis showed ectopic cerebellar cells in the inferior colliculus (Fig. 1B).
Ataxic phenotypes were inherited in a near Mendelian mode (37: 68: 28=1: 1.8: 0.8) and the mutant mice showing the phenotype were all homozygous for the transgene.Mice hemizygous for the transgene in this line were phenotypically normal.On the genetic basis of the recessive mutant trait, coupled with the fact that none of the other lines produced with the same transgenic construct showed the ataxic phenotype, we concluded that the phenotype in line 5 arose from a mutation caused by the insertion of the transgene se-quence into a gene essential for regulating ataxic gait.

Identification of the transgene insertion site
To isolate the transgene insertion site, the restriction digestion patterns of the transgene and its flanking sequences in the genomic DNA of the mutant mice (line 5) were compared with those of the other transgenic line (line 27) generated with the same construct by Southern analysis with probes specific to the transgenic construct.On the assumption that DNA bands unique to the mutant mice contain transgene flanking sequences, while common bands to all of transgenic mice contain transgene sequence only, the DNA fragments unique to the mutant mice were isolated from a gel and were used as templates for inverse PCR after self-ligation.One of the transgene flanking sequences was amplified by PCR from the self-ligated HinfI fractions (~1.7 kb), and confirmed by Southern blot and sequence analysis.The transgene flanking sequence (a' in Fig. 2A) was also amplified from tg/tg genomic DNA by PCR using the primer pair, one specific to the flanking sequence and the other to the transgene sequence (Fig. 2B).The insertion sequence was located as intron 1 of the Unc5h3 gene on chromosome 3 by BLAST search of the ENSEMBL mouse genome database (v.7.3b.3)(Fig. 2A).Based on the sequence information in the database, several sequence-tagged sites (STS) were generated from the flanking region of the transgene (Fig. 2A).While the loci for both a and b were amplified, five consecutive loci (i-1 ~ 5) were not amplified in PCR reactions using genomic DNA from tg/tg mice, suggesting that these mice carry a deletion between a and b (Fig. 2B).The other transgene flanking sequence (b' in Fig. 2A) was also amplified by PCR with another primer pair, one specific to the locus b and the other to the transgene sequence.Sequence analysis of the PCR product revealed a junction between the genomic and transgene sequence.The size of deletion, as estimated by the sequences in the database, was approximately 50 kb and the deleted region corresponds to part of intron 1 of the Unc5h3 gene (approximately 210 kb).Thus, the phenotype of this mutant mouse may be attributable to the transgene insertion and concomitant deletion in intron 1 of the Unc5h3 gene.Alteration in the genomic structure of Unc5h3 was confirmed by pulse-field gel electrophoresis of genomic DNA digested with restriction enzymes and subsequent Southern blot analysis with probes specific to Unc5h3 cDNA (data not shown).This mutation was designated as Unc5h3 rcmTgN(Tert)5Hwl , in accordance with standard nomenclature rules.

Interrupted expression of Unc5h3 gene in many of tissues
To examine whether the transgene insertion mutation could affect expression of the Unc5h3 gene, the mRNA level of the Unc5h3 gene was measured by Northern blot analysis and RT-PCR.While transcripts of the Unc5h3 gene (9.3 kb) were detected in the brain of hemizygous neonates (P7) and in the cerebellum of wildtype adult mice, they were not detected in the homozygous littermates of those mice (Fig. 3C).RT-PCR analysis was also performed to detect the expression of Unc5h3 gene semi-quantitatively in adult mice (Fig. 3A) and neonates (Fig. 3B).Although expressions of the Unc5h3 gene were comparable between wildtype and tg/+ mice in most tissues, they were significantly reduced or undetectable in the cerebellum, lung, spleen and ovary of tg/tg mice.Interestingly, the expression of the Unc5h3 gene was not reduced in the testis of tg/tg mice (Fig. 3A).However, we additionally observed mild defects in splicing by amplifying the sequences within intron 1 of the Unc5h3 gene by RT-PCR from the testis of tg/tg mice which were undetectable in wildtype mice (Fig. 3D).Levels of Unc5h3 expression in tg/tg neonates (P7) were also significantly reduced or undetectable compared with tg/+ littermates.Thus transgenic insertion into intron 1 of the Unc5h3 gene interrupts its expression in most tissues except the testis.

Behavioral analysis of mutant mice
The behavioral phenotypes of homozygous mutant mice were basically similar to those of Unc5h3 rcm mutants described previously [8,21].The swaying gait of the tg/tg mice was identified clearly from 3 weeks of age.Their forward movement is interrupted by swaying to the right or left side (Fig. 4A).However, one of the prominent phenotypes of the tg/tg mice is hyperactivity.They tend to circle in their home cages very frequently and quickly.We examined their spontaneous locomotor activities by the open-field test.Mice were placed in a Plexiglas chamber and their movement was video-tracked for 2 h.Tg/tg mice showed not only extensive ambulatory activity but also circling activity in the chamber (Fig. 4B).The locomotor activity of tg/tg mice, as represented by travel distance, was about 2-4 times higher than those of wildtype and tg/+ (Fig. 4C).Although the average locomotor activity of tg/+ is slightly higher than wildtype, it is not statistically significant.Both tg/+ and wildtype mice exhibited normal locomotor habituation over time.However, tg/tg mice did not show the usual pattern of habituation.Their locomotor activity did not decrease within the 2 h of observation.

Mutant mice show growth retardation
The homozygous mutants are smaller in size and their body weights are 10-30% lighter (P<0.05)than wildtype mice, starting at 10 or 17 days after birth (female or male, respectively) up to 22 months of age, the last time point examined (Fig. 5A).In contrast to the previously described Unc5h3 rcm mutants, which recover their body weight by 5-6 weeks of age [8], our mutant mice did not recover body weight up to 22 months.Major organs (i.e.lung, heart, spleen, kidney, liver and testis) of the tg/tg adult mice except for the brain were similar to those of wildtype mice histologically and in their gross morphology (data not shown).However, fat pads (pericardial, perirenal, mesenteric, and epididymal fat depots) were very small in both male and female tg/ tg mice compared with tg/+ and wildtype mice.The mass of epididymal fat pads of the mutant mice was approximately 20% of that in wildtype and tg/+ mice (Fig. 5B and Fig. 5D).The tg/tg mice do not have problems in food and water intake.Rather, they eat and drink more food and water than wildtype mice (Fig. 5C).They do not seem to be diabetic as their blood glucose levels rose showing the same pattern of wildtype littermates in the fasting blood glucose tolerance test (data not shown).These data suggest that the retarded growth and lower body weight of tg/tg mice do not result from defects of organ development, food and water intake or diabetes but from reduced accumulation of fat tissues.

Discussion
There are two reported mutations in the Unc5h3 gene: one is spontaneous mutation [21], and the other is a transgenic insertional mutation [1].The former mutation is a tandem duplication of an exon encoding amino acids 763-818, and the latter is a transgenic insertion and concomitant deletion of an exon.Here we have described an additional Unc5h3 mutation that was the result of a transgenic insertion causing a deletion in intron 1 of the Unc5h3 gene.While the spontaneous mutant mice express normal levels of Unc5h3 mRNA, both previously reported and our transgenic insertional mutants rescinded expression of the Unc5h3 gene in most tissues.All these mutants show ataxia and hyperactivity due to defects in cerebellar development.
UNC5 family receptors transmit repulsive migratory signals to neurons and thus guide neurons to locate in a proper position, which is critical in the development of functional neuronal circuits [14,19,23,29].The uncoordinated movement of the Unc5h3 mutant mice may result from improper control of motor neurons due to defects in the formation of neuronal circuits in the cerebellum.The small cerebellum of the homozygous mutant mice may indicate a role for UNC5H3 in cerebellar cell proliferation or survival during development.These cerebellar mutant mice showed higher spontaneous activity and a defect in habituation in a novel environment.These phenotypes are similar to those of rats lesioned in the cerebellar vermis [5].The cerebellar vermis has been shown to be involved in motor control, attention capabilities and emotional behavior.Thus our mouse model supports the idea that the cerebellum is important not only for motor control but also for emotional regulation.
The phenotype of reduced fat pads in pericardial, perirenal, mesenteric, and epididymal fat depots may result from the higher energy consumption of the mutant mice due to the hyper ambulatory activity.However, we cannot rule out the possibility that the defective brain function affects the orexigenic phenotype, fat metabolism and adipocyte proliferation/ differentiation of the mutant mice.These possibilities need to be further studied to evaluate the novel lean phenotype of the UNC5h3 mutant mice.
The transgene insertion creating a deletion in intron 1 of the Unc5h3 gene inhibited the expression of Unc5h3 in many tissues including the cerebellum.There are two possible mechanisms by which the transgenic insertional mutation in the intron influences the expression of the Unc5H3 gene.First, transgene insertion itself can interfere with the expression of an endogenous gene.It has been reported that insertion of a Neomycin-resistant gene cassette into an intron often affects the expression of an endogenous gene and results in the reduction or complete inactivation of the gene [13,16,32].The presence of the pLoxPneo gene in intron 10 was found to block normal splicing of the Fgfr3 gene, which lead to phenotypes similar to those of FGFR3-null mice [32].Insertion of PGKneo into intron 2 of the Pafah1b1 gene results in a partial lossof-function allele [13].Although the precise mechanism is largely unknown, we hypothesize that transgenic expression might interfere with splicing and/or transcription elongation, reducing the expression of an endogenous gene.
We observed mild defects in splicing by analyzing the precursor transcript of the Unc5h3 gene by RT-PCR.The sequences within intron 1 of the Unc5h3 gene were undetectable by RT-PCR in the testis of wildtype mice, probably due to immediate splicing followed by transcription.However, they were amplified by RT-PCR from the transgenic allele (Fig. 3D) indicating some of the precursor transcripts derived from transgenic allele were not normally spliced.Despite the presence of immature transcript in the testis of tg/+ and tg/tg mice, the level of mature transcript in tg/tg mice, as represented by RT-PCR, was comparable with that in wildtype mice (Fig. 3A).We postulate that transgene insertion may only moderately delay or inhibit the splicing of Unc5h3 transcripts in the testis.Defects in splicing were not observed in other tissues including the cerebellum.Therefore, other mechanisms may exist which inhibit transcription tissue-specifically after transgene insertion.
Additionally, deletion of an intron sequence that may contain regulatory elements might interfere with transcription initiation or elongation in a tissue-specific manner.Cis-elements within intron sequences play important roles in transcription regulation.Regulation of the cardiac homeobox gene CSX1 is mediated by an enhancer element in the first intron [26].There is substantial evidence for the involvement of the first intron in the tissue-specific expression of the alpha 1 (I) collagen gene [6,15].A negative regulatory element was identified and characterized within the first intron of the epidermal growth factor receptor gene [33].The region in the first exon/intron of rat carnitine palmitoyltransferase I beta is involved in enhancement of basal transcription [31].Parts of the intron sequence are also critical for the basic promoter function of the Cx31 gene [9].Since there are many putative transcription factor binding sites within the deleted region in intron 1 of the Unc5h3 gene, which includes CCAAT, OCT1, GATA, PAX, NKX, HOX, MYT1 and CREB binding sites, the transcription of the Unc5h3 gene could be obstructed by the absence of some of the those regulatory elements.The possible roles of the intron sequence could be investigated by generating intronspecific knockout mice.
In conclusion, the present study suggests that developmental defects in the cerebellum result in not only uncoordinated motion but also spontaneous hyper motor activity and defects in habituation in a new environment.It also shows deleterious effects of transgene insertion even into the intron sequence of an endogenous gene.Transgenic insertion itself and/or its concomitant deletion of some regulatory elements in a non-coding region may interfere with the normal expression of an endogenous gene.Further study will be required to identify the precise mechanism by which transgenic insertion/deletion mutation interrupts the expression of Unc5h3.

Fig. 1 .Fig. 2 .
Fig. 1.Gross morphology and histology of the homozygous mutant (tg/tg) mice.(A) Inferior and superior colliculus of the tg/tg mouse are protruded posteriorly compared with those of the wildtype mouse as indicated by the arrow.(B) The size of the tg/tg cerebellum is smaller than that of the wildtype mouse.(C) Ectopic cerebellar cortical tissues are lined along the dorsal aspect of the inferior colliculus of the tg/tg mouse.Cbl, cerebellum; Mes, mesencephalon; IC, inferior colliculus.

Fig. 3 .
Fig. 3. Unc5h3 gene expressions in the tissues of wildtype (WT), hemizygous (tg/+) and homozygous (tg/ tg) mutant mice.RT-PCR analysis of Unc5h3 transcripts in the tissues of (A) adult and (B) neonate (P7) mice.Gapdh was used as an internal control.(C) Northern blot analysis of Unc5h3 expression in the cerebellum of adult mice and in the brain of neonates.The Unc5h3 transcripts were expressed only in the WT and tg/+ cerebellum or brain as indicated by the arrow.(D) RT-PCR analysis of the precursor transcripts of the Unc5h3 gene in the testis.The sequences in intron 1 (loci a and b) were amplified from the testis cDNA.

Fig. 5 .
Fig. 5. Growth curves and food/water intake.(A) Growth curves of wildtype (WT) male (□, n=22) and female (○, N=19) mice (dotted line) and of homozygous transgenic (tg/tg) male (■, N=5) and female (•, N=5) mice (solid line) based on the body weights measured from 10 days of age to 95 days.Representative WT and tg/ tg mice of 22 months age are shown in the inserted picture.(B) The body and the epididymal fat weight of wildtype (N=9), tg/+ (N=8) and tg/tg (N=3) mice (22 month old males) were compared.Tg/tg mice accumulate significantly smaller epididymal fat pads than wildtype and tg/+ mice.*: P<0.05, Student's t-test.(C) Daily food and water intake of adult WT, tg/+ and tg/tg mice.Average daily intake of the tg/tg mice was significantly more than that of wildtype mice.*: P<0.05, Student's t-test.(D) Photographs of epididymal fat pads of WT and tg/tg mice (22 month old male).