Cloning, functional study and comparative mapping of Luzp2 to mouse Chromosome 7 and human Chromosome 11p13–11p14

Introduction

The morphological specific-locus test performed at the Oak Ridge National Laboratory has, over the course of five decades, generated a large collection of radiation- or chemically induced deletion mutations surrounding the pink-eyed dilution locus (p) on mouse Chr 7 (Russell 1951; Rinchik and Russell 1990; Russell et al. 1995; Johnson et al. 1995). These deletions have been valuable experimental tools in cloning and characterizing genes correlated with specific deletion phenotypes.

Complementation and molecular analyses of p-region deletion mutations identified 19 deletions whose proximal deletion breakpoints fall within the interval between growth arrest specific 2 (Gas2) and hect [homologous to the E6-AP (UBE3A) carboxyl terminus] domain and RCC1 (CHC1)-like domain (RLD) 2 (Herc2) (Johnson et al. 1995; Russell et al. 1995; Ji et al. 1999). A functional unit (l7Rl1) that is essential for early embryonic survival also maps to this interval (Russell et al. 1995; Wu et al. 2000). In order to study biological processes specified by genes within this interval, and before the complete public human and mouse genomic sequences were available, we sought to characterize molecularly the chromosomal region between Gas2 and Herc2.

We took advantage of the previous molecular analysis of one p-region deletion, Del(p)^8FDFoD, to provide an initial access point to this region (Fig. 1A). The proximal breakpoint p ^8FDFoD had previously been mapped by genetic complementation tests to the Gas2–Herc2 interval, whereas the distal breakpoint was known to be within the Gabrb3 gene (Culiat et al. 1993; Johnson et al. 1995; Russell et al. 1995). Analysis of Gabrb3 expression in the brain of a series of p-deletion compound heterozygotes showed the presence of a 6.5-kb mutant transcript arising from the p ^8FDFoD chromosome, instead of the 5.5-kb wild-type Gabrb3 transcript (Culiat et al. 1993), making it likely that the mutant transcript was the result of a fusion of Gabrb3, at the distal deletion breakpoint, to a gene mapping at the proximal p ^8FDFoD breakpoint. We report here the cloning of this fusion transcript, and subsequent expression and mutational analyses of the novel proximal gene, leucine zipper protein-2 (Luzp2), mapping within the Gas2–Herc2 region of mouse Chr 7. We mapped the human homolog, LUZP2, to Chr 11p13–11p14 and then more finely mapped this gene within the common region for chromosomal deletions associated with WAGR syndrome.

Figure 1

A) Schematic deletion map of selected p-deletion mutations (adapted from Johnson et al. 1995; Russell et al. 1995; Rinchik et al. 1995). The heavy black line represents the chromosome segment surrounding the p gene. The horizontal lines beneath the chromosome represent the chromosome segments deleted in the corresponding p-deletion mutations. Genes and molecular markers are shown above the chromosome. ■ represents the gene at the proximal end of the p ^8FDFoD deletion breakpoint. □ represents the Gabrb3 gene at the distal end of the p ^8FDFoD deletion breakpoint. The map is not drawn to scale. B) Summary of the primers used in the RACE analyses as well as the cDNA fragments obtained from RACE. Details may be found in Materials and methods. C) Mapping of 5′RACE-a. Genomic DNA, prepared from mice of the indicated genotypes, was digested with PstI, electrophoresed, blotted, and hybridized with probe a. H is an F₁ hybrid of 101/R1 × C3Hf/R1, used as control for p-deletion mutations. “Spt” stands for M. spretus, and “+” represents the wild-type chromosome from Mus musculus. Note the Mus musculus-specific band (M band) is deleted in p ^46DFiOD/Spt, but is not deleted in p ^8FDFo/Spt, or in p ^3FR60Lg. D) Mapping of 3′RACE-c. Genomic DNA, prepared from animals of the indicated genotypes, was digested with TaqI and hybridized with probe 3′RACE-g. Note the Mus musculus-specific band (M band) is deleted in p ^8FDFoD/Spt, but is not deleted in p ^226THO-I/Spt, p ^12R30Lb/Spt, and p ^3FR60Lg.

Materials and methods

Mice

p-deletion mutations used in this study were generated at Oak Ridge National Laboratory (Russell et al. 1995) and are currently maintained as p ^l/p ^7R75M heterozygotes. p ^l generically represents any lethal p-deletion mutation, and p ^7R75M is a radiation-induced viable allele of p that confers darker coat color than most alleles of p. p ^l/p ^7R75M heterozygotes are lighter in color than are p ^7R75M/p ^7R75M mice. The T (“Tester”) mice used in the rapid amplification of cDNA ends (RACE) experiments are from a randomly mated stock homozygous for recessive alleles at seven visible marker loci (a, nonagouti; Tyrp ^b, brown; p, pink-eyed dilution; Tyr ^c-ch, chinchilla, an allele at the albino [Tyr(c)] locus; Myo5a ^d, dilute; Bmp5 ^se, short ear; Ednrb ^s, piebald spotting).

RACE

The RACE kit (Marathon^™ cDNA amplification kit, Cat.# K1802-1, CLONTECH Laboratories, Inc.) was used according to directions from the manufacturer. Poly-A⁺ brain RNAs from p ^8FDFoD/p ^4THO-II and T-stock E14.5 embryos were used to generate cDNA templates. PCR parameters: 94°C for 4 min; 28 cycles of: 94°C for 30 s, 68°C for 4 min; 4°C, hold. Primers used in RACE are:

m1, 5′-GGGTGCAGTTTTGCTCATCCAGTGGGTATC-3′;

m2, 5′-CTGAGGTCCATCATGCAAGCTGCTGTGGTG-3′;

m3, 5′-CAGGAGTGATGTCACCCC-3′;

m4, 5′-GAAGCCCGATACCCAGATGAGTTAGTGGTC-3′;

m5, 5′-CCAACCCAGATGTTACATCCACCCAGGACT;

m6, 5′-GACTGTTGCTTCCAGGATTCCAGAAGCCAG;

Primers m1 and m2 were designed from the rat Gabrb3 cDNA sequences (accession number NM_017065). Primers m3–m6 were designed from the Luzp2 cDNA sequences (accession number AY164459).

Ribonuclease protection assay (RPA)

The Ribonuclease Protection Assay Kit (Ambion) was used according to manufacturer's instructions. p650 is a 650-bp PstI subfragment of Luzp2 cDNA cloned into a PCR^™ 2.1 vector. ³²P-UTP (800 Ci/mmol) was used in the transcription reaction for radio-labeling the riboprobes, and a 80,000-cpm probe and 10 µg of total RNA were used in each hybridization reaction. The hybridization mixtures were then separated on a polyacrylamide gel and visualized by autoradiography.

Radiation hybrid mapping

The GeneBridge 4 Radiation Hybrid Panel, purchased from Research Genetics in a PCR amplifiable format, consists of 93 radiation hybrid clones from the whole human genome. Human LUZP2 and GAS2 (accession# NM_005256) specific primers were used to screen this radiation hybrid panel. Primer sequences are as follows: LUZP2-For, 5′-CTCGCTCCCAGCGCCT-GCC-3′; LUZP2-Rev, 5′-GACCAGCGCAGGCAG-GAGAG-3′; GAS2-For, 5′-CCTTTTAATTCTTGA-ATGAG-3′; GAS2-Rev, 5′-GTTAATCCCTATTTT-GCTAT-3′.

Mapping of LUZP2 to the WAGR deletion region

Somatic cell hybrid DNAs (GOX1.5.6, SOA54, BAN68, RIX1.4, NYX3.1.5, TRX9.4F4, and ANX6.14), in which only the mutant Chr 11 of the WAGR patients is retained, were obtained from Wendy Bickmore (MRC Human Genetics Unit, Edinburgh, UK). Primers used in PCR amplification of the LUZP2 genomic fragment in this analysis are the LUZP2-For and LUZP2-Rev primers listed above. Primers used in the amplification of the SAA3 genomic fragment used as a control for the testing of deletion of LUZP2 from WAGR patients are designed from the SAA3 amplimer (GDB: 4590133): SAA3-a, 5′-ATGCAAGCAGCATTATCCCTT-3′; SAA3-b, 5′-CTGCTCTGCTTTGGAACTTCTTT-3′.

DNAs from WAGR patient cell cultures (GM05518, GM03808, GM07736, GM06803, GM08785, GM03118A, GM05296, and GM09709) were purchased from the Institute for Medical Research (Camden, NJ). The genomic DNAs were digested with BamHI, electrophoresed, blotted, and hybridized with LUZP2.380 and DN10.400. LUZP2.380 is a 380-bp PstI-EcoRI fragment from the LUZP2 cDNA, and DN10.400 is a 400-bp PstI-EcoRI fragment from the P cDNA (accession # NM_021879). The intensities of the LUZP2-specific band and the P-specific band on the hybridized Southern blot were quantified on a Phosphor Imager, and the ratio of LUZP2:P within each lane was calculated.

Construction of the Luzp2 targeting vector

A 12-kb genomic DNA fragment encompassing exon 1 of Luzp2 was isolated from a 129/Sv mouse genomic library constructed by Edward Michaud, Oak Ridge, Tenn. The β-gal-Neo cassette in the pSAβgal targeting vector (Friedrich and Soriano 1991) was used to replace the 2.2-kb SacII-XbaI Luzp2 genomic fragment including partial sequences of both exon 1 and intron 1. As a result, Luzp2 cDNA sequences from position −663 to position +61 relative to its translation initiation codon were removed in the construct. In addition, there are 2.3-kb and 3.4-kb homology arms 5′ and 3′, respectively, of the β-gal-Neo cassette insertion, in the targeting construct.

Generation and genotyping of Luzp2 mutant mice

The KpnI linearized construct was electroporated into the R1 ES cell line. G418-resistant ES clones were first screened by PCR for homologous recombination at the 5′ end. Primers used were forward 5′-GAGCTGTTAAGGGATCTGGGAAGTC-3′; and reverse 5′-CGACGGGATCCGCCATGTCACAGATCATCA-3′. Targeted ES cells should yield a 2.7-kb product by this PCR amplification, and non-targeted ES cells should have no PCR product. Positive ES cells were then screened by Southern blot hybridization for homologous recombination at the 3′ end. Probe A, a 1-kb XbaI–EcoRI fragment in intron 1 of Luzp2, detects a 7.1-kb EcpRI genomic fragment from the Luzp2 wild-type allele, and detects a 6.3-kb EcoRI genomic from the Luzp2 targeted allele.

Out of 215 G418-resistant clones, 14 correctly targeted clones were identified. ES cells carrying the Luzp2 targeted allele were then microinjected into C57BL/6J blastocysts. Chimeric males were mated with C3Hf/Rl females to obtain Luzp2 ^+/− heterozygotes, which were then intercrossed to obtain homozygous Luzp2 ^+/− mice. Primers used for subsequent genotyping PCR were: ORN484 5′-GTAGGTCAGCCTCTC-3′, ORN498 5′-CGACGGGATC CGCCATGTCACAGATCATCA-3′. ORN484 and ORN498 amplify a 320-bp fragment from the Luzp2 targeted allele, whereas no amplification product should be produced from the wild-type allele. Primers used in testing for Luzp2 transcription in the homozygous Luzp2 ^−/− mice were as follow: ORN457 5′-CTACGCTAGAGACCACTAAC-3′; ORN471 5′-CGTGCTCCGCAGGGTGAGAG-3′. This primer pair should amplify a 283-bp Luzp2 cDNA fragment from nt. 80 to 303.

Results

Cloning of Luzp2 by RACE of the Del(p)^8FDFoD fusion transcript

p ^8FDFoD/p ^4THOII poly A⁺ brain RNA, in which only the 6.5-kb Gabrb3 fusion transcript was detected when probed with Gabrb3 cDNA (Culiat et al. 1993), was used to generate a cDNA template for RACE. Adapter sequences were linked to both ends of the cDNA templates. Nested primers, m1 and m2 in Gabrb3, together with adapter-specific primers AP-1 and AP-2, were designed to amplify the 5′-most Gabrb3 sequences, as well as the novel sequences, in the 5′ end of the fusion transcript (Fig. 1B). Two consecutive PCRs were performed, the first using primers m1 and AP-1, and a second nested PCR using m2 and AP-2. As illustrated in Fig. 1B, a 1500-bp cDNA fragment, 5′RACE-a, was amplified through 5′RACE. Sequence analysis of 5′RACE-a revealed a 250-bp match to the rat Gabrb3 sequences, but the remaining 1250-bp sequences did not match Gabrb3 sequences. Thus, we assumed that the non-homologous sequence to be part of a gene disrupted by the proximal end of the p ^8FDFoD deletion.

Probe a was generated from the novel portion of 5′RACE-a for mapping by using Restriction Fragment Length Variation (RFLV) analyses (Johnson et al. 1989, 1995) (Fig. 1C). Genomic DNA from (Mus musculus/Mus spretus) F₁ hybrids carrying p ^46DFiOD has only the M. spretus-specific band (S band) detected by probe a, whereas the Mus musculus-specific band (M band) is absent, showing that probe a is deleted in the p ^46DFioD mutant chromosome. Probe a is not deleted in p ^8FDFoD or in p ^3FR60Lg, as shown by the presence of the Mus musculus-specific bands in these F₁ hybrid DNAs. Thus, probe a, corresponding to a new gene at the 5′ of the Gabrb3 fusion transcript in p ^8FDFoD, maps to me region between the 5′ deletion breakpoints of p ^46DFiOD and p ^8FDFoD. We named this new gene Luzp2 to reflect the presence of a leucine-zipper motif in the deduced protein (see below).

In order to amplify more 5′ end sequences for the Luzp2 cDNA, additional primers (m3 and m4) were designed from the 5′ end of 5′RACE-a sequences and used in one more round of 5′RACE. An additional 150-bp fragment, 5′RACE-b, was amplified from E14.5 p ^8FDFoD/p ^4THO-II brain cDNA. RT-PCR analysis confirmed that 5′RACE-b is from the Luzp2 transcript, extending further towards the 5′ end than 5′RACE-a (Fig. 1B).

As in 5′RACE, primers (m5 and m6) were designed from the 3′ end of existing Luzp2 sequences and used in 3′RACE to clone wild-type Luzp2 cDNA fragments distal to the breakpoint that generated the fusion transcript (3′RACE-c in Fig. 1B). A 3.8-kb Luzp2 clone was amplified from p ^4THO-II chromosome of E14.5 p ^8FPFoD/p ^4THO-II brain cDNA. The 3′RACE clone maps within the p deletion complex as determined by RFLV analysis, showing that the 3′RACE clone is deleted in p ^8FDFoD, but is present in p ^3FR60Lg (Fig. 1D).

Overall, a combined 5.2-kb Luzp2 cDNA was cloned through RACE. RFLV analyses confirmed that the 5′RACE fragment is present in the p ^8FDF ° ^D mutant chromosome, whereas the 3′RACE fragment is deleted in p ^8FDFoD, demonstrating that Luzp2 is indeed disrupted by the proximal deletion breakpoint of p ^8FDFoD.

Luzp2 is limited to the central nervous system of adult mice

In order to analyze the expression ot Luzp2 in adult mice, probe a was hybridized to a Northern blot containing poly-(A)⁺ mRNAs from adult mouse brain. Two transcripts, 5.2 kb and 4.5 kb in size, were detected in the brain (Fig. 2A).

Figure 2

Analyses of Luzp2 expression patterns in adult mice. A) Northern blot with poly-(A)⁺ RNA, prepared from the brain of C3H/R1 adult mice hybridized with probe a, a Luzp2 cDNA fragment. The lower panel is the same Northern-blot hybridized with a tubulin probe. B) Ribonuclease Protection Assay using total RNA from the indicated mouse tissues hybridized with ³²P-labeled p650 sense or antisense probes-actin was added to every reaction as a control.

In order to determine the orientation of Luzp2 transcription and to see whether the Luzp2 transcripts could be detected in other tissues, a ribonuclease protection assay (RPA) was performed by hybridizing Luzp2 antisense and sense probes to RNAs from various adult tissues. The Luzp2 (presumed) antisense probe detected the Luzp2 cDNA fragment in brain and spinal cord, but not in any other tissues tested (Fig. 2B). There was no hybridization of the Luzp2 sense probe, leading to the conclusion that Luzp2 is transcribed in a centromere-to-telomere orientation.

Analysis of Luzp2 cDNA sequences

The cDNA sequence of Luzp2 was obtained by sequencing 5′RACE and 3′RACE Luzp2 cDNA fragments (GenBank accession AY164459). The Luzp2 cDNA is 5184 bp in length and has a 1037-bp continuous open reading frame, with a 711-bp 5′ untranslated region (UTR) and a 3521-bp 3′ UTR (Fig. 3). A consensus polyadenylation signal AATAAA is located at nt. 5158. A BLAST search indicated that Luzp2 is a novel gene with many EST hits from various parts of the brain (e.g., cerebellum, brain stem, olfactory bulb, hypothalamus, cortex, amygdala, basal ganglia, pineal gland, striatum, hippocampus). Protein sequence analysis of Luzp2 identified a leucine-zipper motif at aa. 164–192; this sequence is characterized by the presence of leucines at every seventh residue over some 28 to 35 amino acids (Fig. 3).

Figure 3

Nucleotide sequences and deduced amino acid sequences of Luzp2. The translation initiation codon and the stop codon are underlined. The * marks the stop codon. Leucine residues within the leucine zipper motif are boxed. The arrowhead indicates the deletion breakpoint of Luzp2 in the p ^8FDFoD mutation.

LUZP2 maps to human 11p14, within the region associated with mental retardation in WAGR patients

The human homolog of mouse Luzp2 was cloned and mapped to see whether the human map position (and/or association with any mapped human diseases) might suggest a function for the brain and spinal cords-specific Luzp2.

With mouse Luzp2 cDNA as a probe, a 1532-bp cDNA clone for the human LUZP2 homolog (GenBank accession AY164460) was cloned from a human brain cDNA library. The 1532-bp partial LUZP2 cDNA sequence has 73.8% sequence identity to the mouse Luzp2 cDNA from nt. 529 to nt. 2061, encompassing the protein-coding region of Luzp2.

Primers were designed from these cDNA sequences and were used to map LUZP2 on the GeneBridge 4 radiation hybrid panel. Analysis of the mapping-panel data with RH mapping tools at http://www-genome.wi.mit.edu/cgi-bin/contig/rhmapper.pl.fordfay placed LUZP2 on human Chr 11 between STS marker WI-10394 and CHLC.GATA25B04 (Fig. 4). As a reference, primers were designed from GAS2 sequences (NM_305256) and used to map GAS2 to the interval between WI-3548 and WI-13884.

Figure 4

Homology map of human 11p15.1-p12 to mouse Chrs 7 and 2. The MIT Radiation Hybrid map and cytogenetic map of the human 11p15.1-11p12 are listed at the right side of the figure. Gene symbols are written in boldface, STS markers are written in plain text. Segments of mouse Chrs 7 and 2 are shown at the left side of the figure.

As shown in Fig. 4, the cytogenetic position of LUZP2 was estimated by comparing the MIT Radiation Hybrid map with the genemap at http://www.ncbi.nlm.nih.gov/genemap99 . WI-9500 had been mapped to 11p14–11p13, and WI-13781 to 11p13. Since markers WI-10394 and CHLC.GATA25B04 are both mapped within the interval between WI-9500 and WI-13781 in the Radiation Hybrid map, LUZP2 should be located within 11p13-11p14.

Consistent with the RH mapping data, a more recent BLAST search with LUZP2 partial cDNA sequences has mapped the gene to a human Chr 11 BAC RP11-72C9 (accession number AC067902), which is located at 11p14.3, 26.0 MB from the centromere of human Chr 11 in the human genome assembly (the June 2002 freeze, http://genome.ucsc.edu/ ). Since GAS2 and BDNF have been located to 24.1 MB and 29.1 MB respectively on Chr 11, LUZP2 is located between those two genes (Fig. 4).

LUZP2 is deleted in several WAGR patients

In humans, heterozygous deletion of 11p13-14 predisposes the affected individual to develop WAGR syndrome, associated with Wilms tumor of the kidney, aniridia, genitourinary abnormalities, and various degrees of mental retardation (Russell and Weisskopf 1986). By analyzing the Chr 11 deletion breakpoints from a number of WAGR-related patients, Fantes et al. (1995) proposed that a gene(s) predisposing to mental retardation might reside distal to BDNF in 11p14.

The gross map position of LUZP2 in Chr 11p13–11p14 suggested that the gene could be associated with WAGR syndrome, and the availability of genomic DNA from WAGR patients offered a tool for finer mapping of this gene. Therefore, we screened for the presence or absence of LUZP2 in DNAs from WAGR patients. Somatic cell hybrid DNAs, which were used in constructing a WAGR patient deletion map and kindly shared by Wendy Bickmore, were used in PCR analysis with LUZP2-specific primers. As shown in Fig. 5A, the LUZP2 genomic fragment was amplified from all samples tested except NYMI, showing that LUZP2 is deleted in one mentally retarded patient whose deletion extends most distally among the ones tested.

Figure 5

Mapping of LUZP2 to WAGR-related deletions. A DNAs from the indicated somatic-cell hybrids were used in PCR analysis with LUZP2-specific primers, LUZP2-For, and LUZP2-Rev. Those somatic-cell hybrids are derived from the cell cultures of WAGR-related patients, and only the mutant human Chr 11 is retained in those hybrids. The 225-bp LUZP2 genomic DNA fragment was amplified from all of the samples except NYMI. 4A23 is hamster DNA. The serum amyloid A three (SAA3) gene maps to 11p15 and has been shown to be present in the patients tested here (Fantes et al. 1995). As a positive control, SAA3-specific primers were used in the PCR analysis. B) BamHI-digested genomic DNAs from the cell cultures of the indicated patients, hybridized with LUZP2.380 and DN10.400. The last two lanes are hybridizations of human placenta DNA with LUZP2.380 or DN10.400. The ratios of the intensities of LUZP2 band:P band are listed below each lane.

Another deletion map of the WAGR region was generated that included 13 WAGR-related deletions (Gessler et al. 1989). Human DNAs used in constructing that map were also tested for the presence of LUZP2. Since the DNAs are from WAGR-related human cell cultures, which are heterozygous for the corresponding deletions, quantitative Southern-blot analysis was used to estimate LUZP2 gene copy number in those DNAs. As control for densitometry, DN10.400, a probe from the P gene that maps to human 15q13, was also added in the hybridization reactions (Fig. 5B). The intensities of the LUZP2-specific band and the P-specific band were quantified on a Phosphor Imager, and the ratio of LUZP2:P within each lane is listed below the individual lanes. Assuming that the P DNA is presents in the same amount in all of the DNA samples, the quantities of LUZP2 DNA in GM05518 and GM03808 are about one-half those in GM07736, GM06803, GM08785, and GM03118A. Therefore, LUZP2 appears to be deleted in the mutant chromosomes of GM05518 and GM03808, the two mentally retarded patients whose deletions extend the most distally. LUZP2 is not deleted in the other mentally retarded patients with shorter deletions distally, nor is it deleted in WA patients with normal intelligence.

Generation of a Luzp2 mutant mouse model

Because a peri-implantation lethality mutation, l71Rl, is located between p and Luzp2 on mouse Chr 7 (Wu et al. 2000), mice homozygous for the existing p deletions die early in gestation and thus are unavailable for studying the biological function of Luzp2. A Luzp2 knockout mouse model was therefore generated through homologous recombination in embryonic stem (ES) cells.

Using Luzp2 cDNA as a probe, a Luzp2-positive genomic clone was obtained from a phage genomic library. Comparison of the sequences of the genomic clone to the Luzp2 cDNA identified the first exon of Luzp2 to be from nt. 1 to 774, including the ATG start codon at nt 713. We therefore decided to disrupt exon1 of Luzp2 in the targeted mice (Fig. 6A). Fig. 6B shows the Southern-blot hybridization used to screen for homologous recombination between the Luzp2 wild-type allele and the targeting construct. Positive ES cells were microinjected into C57B1/6 blastocysts to produce chimeric males, which were then bred to C3Hf females. Male and female offspring heterozygous for the Luzp2 ^+/− targeted allele were intercrossed to produce homozygous Luzp2 ^+/− mutant mice.

Figure 6

A) Generation of a targeted allele at the mouse Luzp2 locus. Homologous recombination between the wild-type allele and the targeting vector results in the replacement of some exon1 sequences with the _-gal-Neo selection cassette. Probe A is a 1-kb XbaI-EcoRI fragment of Luzp2 used in screening for the targeted allele by Southern-blot hybridization. R is the abbreviation for EcoRI. B) Screening for ES clones carrying a Luzp2-targeted allele. DNAs from G418-resistant ES clones were digested with EcoRI, electrophoresed, blotted and hybridized with probe A, which identifies EcoRI fragments of both 7.1 kb (wild-type allele) and 6.3 kb (targeted allele). C) RT-PCR analysis of Luzp2 ^−/− mutant mice. Brain RNAs were prepared from mice of indicated genotypes, reverse-transcribed, and amplified by PCR with Luzp2-specific primers ORN457 and ORN471. The expected amplification product is 223 bp. As a control, the same RT reactions were also amplified with β-actin primers.

Characterization of Luzp2^+/− mutant mice

Luzp2 ^−/− mice are viable, fertile, and have no apparent abnormalities. RT-PCR analysis was used to examine Luzp2 ^−/− expression in the brains of homozygous knockout mice. Although a Luzp2-specific transcript was amplified from the wild-type control mice, no Luzp2 cDNA fragment was amplified from the brains of four individual Luzp2 ^−/− mice, confirming that Luzp2 transcripts are absent in the mutant mice (Fig. 6C).

Discussion

Luzp2 was cloned based on the hypothesis that the size-altered Gabrb3 transcript in p ^8FDFoD/p ^4THO-II compound heterozygotes results from the fusion of two separate genes located at the proximal and distal ends of the p ^8FDF ° ^D deletion. The deletion in the p ^8FDF ° ^D chromosome removes 3.8 kb of 3′ sequence of Luzp2 cDNA at its proximal end, and the Gabrb3 promoter plus 320 bp of Gabrb3 5′ end cDNA sequences (accession NM_008071) at its distal end. Transcription initiated from the Luzp2 promoter continues into the remaining Gabrb3 sequences to produce the 6.5-kb Luzp2/Gabrb3 fusion transcript. The fusion protein translated from the Luzp2/Gabrb3 transcript is composed of 253 aa out of 345 aa Luzp2 protein sequences at the N-terminus and in-frame 394 aa out of 473 aa Gabrb3 protein sequences at the C-terminus. Since p ^8FDF ° ^D/+ heterozygotes appear to be normal and fertile, the fusion protein does not seem to have any dominant effect in the affected mice. On the other hand, p ^8FDFoD/ p ^8FDF ° ^D homozygotes exhibit cleft palate and die prenatally, indicating the Gabrb3 function is disrupted in p ^8FDF ° ^D. The loss of function of Gabrb3 in p ^8FDF ° ^D may be explained by the missing N-terminal 79-aa Gabrb3 protein sequences or by the fact that, since the fusion transcript is under the control of the Luzp2 promoter, the fusion protein may not be present in the right place or right time for Gabrb3 function. The leucine-zipper-motif is still retained in the fusion protein, but it is not clear whether or not the fusion protein can at least partially replace the function of Luzp2 since we don't yet know the exact function of Luzp2.

Luzp2 is a novel gene that is expressed only in brain and spinal cord among all tissues tested. Radiation hybrid mapping placed LUZP2 on human Chr 11p13–11p14, within the critical interval for WAGR syndrome in humans. Furthermore, LUZP2 maps within the deletion interval of 3 out of 15 WAGR patients tested.

A genetic origin for the MR component of WAGR syndrome was first suggested by Russell and Weisskopf (1986) from their work in which 43 WA patients were analyzed for their intelligence and the sizes of associated chromosomal deletions. Overall, 67% the WA patients examined were mentally retarded (MR), but patients with smaller deletions were significantly more likely to have normal intelligence.

Fantes et al. (1995) characterized the positions of the chromosomal deletion breakpoints of several WAGR patients versus WAG patients. The sizes of the chromosomal deletions are not always concordant with the degree of MR reported for those patients. In general, however, patients with deletions extending further distally than the BDNF gene tend to have more severe mental retardation. Therefore, Fantes et al. proposed that a gene(s) predisposing to MR may reside distal (telomeric) to BDNF.

We have mapped LUZP2 to the WAGR region deletion map composed by Fantes et al. (1995) and to another map constructed by Gessler et al. (1989). LUZP2 maps within deletions found in three mentally retarded patients, but it does not map within the other, less distally extending deletions also found in mentally retarded patients (Fig. 5). Therefore, if the MR in WAGR is caused by mutations in a single gene, that gene must map between BDNF and LUZP2, which excludes LUZP2 as a candidate.

However, a review of 37 reports of aniridia patients with partial deletions of Chr 11p concluded that MR in those patients is almost constantly present but is highly variable, with some patients being reported as normal, or borderline, and others as severely affected (Turleau et al. 1984). This suggests that MR may be a result of the compounding effects by mutations in several genes in a contiguous gene syndrome, and/or modified by one or many background genes in different individuals.

So far, three brain-specific genes have been mapped within the WAGR critical interval, and each may contribute to MR in some yet undefined way. PAX6 is the gene responsible for causing the aniridia component of WAGR syndrome in humans (Jordan et al. 1992). Severe central nervous system defects and absence of eyes are observed in human PAX6 homozygotes (Glaser et al. 1994). The second brain-specific gene deleted in WAGR patients is BDNF (Brain-Derived Neurotrophic Factor), a member of the NGF (nerve growth factor) family of polypeptide neurotrophins (Leibrock et al. 1989). Mapping of BDNF relative to the deletion interval in some WAGR patients indicated that, while not all MR patients are deleted for BDNF, there appears to be a correlation between BDNF dosage and the severity of MR phenotype (Hanson et al. 1992). The third gene mapped to the WAGR critical interval is 239FB. This gene maps to 11p13 between PAX6 and BDNF, is predominantly expressed in fetal brain, and is, therefore, considered a candidate gene for causing MR.

LUZP2 is the fourth brain- and spinal cord-specific gene mapped to this WAGR interval. It is worth considering that mutations in LUZP2, together with the above three brain-specific genes, may contribute to the severity of mental retardation in WAGR patients.

To examine the function of Luzp2 in mice and to determine whether the loss of one copy of the gene may contribute to the mental retardation observed in WAGR patients, we generated mice both heterozygous and homozygous for a null mutation by gene targeting. Southern-blot hybridization analysis indicated that the targeting construct has been integrated into the Luzp2 locus through homologous recombination (Fig. 6B). In addition, no Luzp2 transcript could be detected in the brain of Luzp2 ^−/− mice (Fig. 6C), confirming that we have generated a null mutation in Luzp2. Although no obvious behavioral abnormality was observed in Luzp2 ^−/− homozygous knockout mice, we plan to subject these mice to a rigorous set of behavioral tests designed to detect abnormalities in learning and memory, increased anxiety level, behavioral despair, or locomotor activity, in hopes of modeling any cognitive deficits analogous to human phenotypes characteristic of WAGR.

Radiation hybrid mapping placed LUZP2 in human 11p13-11p14. Previously, the human homologs of Gas2 and Myod1, two genes that are proximal to Luzp2 on mouse Chr 7, were mapped to human 11p14.3-11p15.1 (Collavin et al. 1998; Stubbs et al. 1994). Therefore, the mouse chromosome region between Luzp2 and Myod1 is homologous to human 11p13-p15.1. Herc2, however, which is distal to Luzp2 on mouse Chr 7, maps to human 15q13 (Ji et al. 1999). Therefore, the mouse Chr 7 segment homologous to human 11p13 ends somewhere between Luzp2 and Herc2, which are less than 1 Mb apart in mouse (Wu et al. 2000).

In human Chr 11, the segment from WT1 to BDNF (Fantes et al. 1995) is homologous to the chromosome segment between Wt1 and Bdnf in mouse Chr 2. Therefore, the mouse Chr 7 homologous region must start somewhere distal to BDNF in human 11p13. The p-deletion mutations that extend proximally to Luzp2 are, therefore, good genetic “reagents” to study human diseases mapping to 11p13-p15.1 (see, for example, Rinchik et al. 2002). Specifically, further studies could be focused on identifying additional MR candidate genes in the Luzp2 region of mouse Chr 7.

Previously, a peri-implantation lethal locus, l71Rl, was mapped to the interval between Gas-2 and Herc2. Since Luzp2 is not deleted in some of the l71Rl deletion mutations, Luzp2 is unlikely to be the candidate gene for causing the peri-implantation lethality in the mutant embryos. However, Luzp2 provided an access for constructing a physical map encompassing the l71Rl minimum candidate interval (Wu et al. 2000), which is a valuable resource for evaluating candidate genes for l71Rl, as well as for studying the function of other genes within that chromosome interval.