Identification of the gene that, when mutated, causes the human obesity syndrome BBS4.

Bardet-Biedl syndrome (BBS, MIM 209900) is a heterogeneous autosomal recessive disorder characterized by obesity, pigmentary retinopathy, polydactyly, renal malformations, mental retardation, and hypogenitalism. The disorder is also associated with diabetes mellitus, hypertension, and congenital heart disease. Six distinct BBS loci map to 11q13 (BBS1), 16q21 (BBS2), 3p13-p12 (BBS3), 15q22.3-q23 (BBS4), 2q31 (BBS5), and 20p12 (BBS6). Although BBS is rare in the general population (<1/100,000), there is considerable interest in identifying the genes causing BBS because components of the phenotype, such as obesity and diabetes, are common. We and others have demonstrated that BBS6 is caused by mutations in the gene MKKS (refs. 12,13), mutation of which also causes McKusick-Kaufman syndrome (hydrometrocolpos, post-axial polydactyly, and congenital heart defects). MKKS has sequence homology to the alpha subunit of a prokaryotic chaperonin in the thermosome Thermoplasma acidophilum. We recently identified a novel gene that causes BBS2. The BBS2 protein has no significant similarity to other chaperonins or known proteins. Here we report the positional cloning and identification of mutations in BBS patients in a novel gene designated BBS4.

Positional cloning of a novel gene on chromosome 16q causing Bardet-Biedl syndrome (BBS2).

Bardet-Biedl syndrome (BBS) is a genetically heterogeneous autosomal recessive disorder with the primary clinical features of obesity, pigmented retinopathy, polydactyly, hypogenitalism, mental retardation and renal anomalies. Associated features of the disorder include diabetes mellitus, hypertension and congenital heart disease. There are six known BBS loci, mapping to chromosomes 2, 3, 11, 15, 16 and 20. The BBS2 locus was initially mapped to an 18 cM interval on chromosome 16q21 with a large inbred Bedouin kindred. Further analysis of the Bedouin population allowed for the fine mapping of this locus to a 2 cM region distal to marker D16S408. Physical mapping and sequence analysis of this region resulted in the identification of a number of known genes and expressed sequence tag clusters. Mutation screening of a novel gene (BBS2) with a wide pattern of tissue expression revealed homozygous mutations in two inbred pedigrees, including the large Bedouin kindred used to initially identify the BBS2 locus. In addition, mutations were found in three of 18 unrelated BBS probands from small nuclear families.

Cloning and characterization of T-cell lymphoma invasion and metastasis 2 (TIAM2), a novel guanine nucleotide exchange factor related to TIAM1.

TIAM1 is a guanine nucleotide exchange factor that was identified in a screen for genes that increase the invasiveness of T lymphoma cell lines (Habets et al., 1994, Cell 77(4): 537-549). We have identified a gene, T-cell lymphoma invasion and metastasis 2 (HGMW-approved symbol TIAM2), with significant identity to the carboxyl-terminal region of the TIAM1 and mapped it to 6q25. TIAM2 is expressed as an approximately 3.3-kb transcript in cerebrum and as an approximately 4.4-kb transcript in the cerebellum and testis. The approximately 4. 4-kb message encodes a longer form of the approximately 3.3-kb mRNA predicted protein, and both contain homology to the Dbl-homologous region (70%) and Pleckstrin-homologous (54%) regions of TIAM1. We have purified TIAM2 and shown it to have GDP-GTP exchange activity. In situ hybridizations demonstrate TIAM2 expression in the E13.5 telencephalon of mouse embryos and in the cerebral cortex, hippocampus, and ependyma of adult mouse brains.

The cloning and developmental expression of unconventional myosin IXA (MYO9A) a gene in the Bardet-Biedl syndrome (BBS4) region at chromosome 15q22-q23.

Bardet-Biedl Syndrome (BBS) is a heterogeneous, autosomal recessive disorder characterized by mental retardation, obesity, retinitis pigmentosa, syndactyly and/or polydactyly, short stature, and hypogenitalism and is caused by mutations at a number of distinct loci. Using a positional cloning approach for identifying the BBS4 (chromosome 15) gene, we identified and cloned an unconventional myosin gene, myosin IXA (HGMW-approved symbol MYO9A). Since mutations in unconventional myosins are known to cause several human diseases, and since mutations of unconventional myosin VIIa cause retinal degeneration, we evaluated myosin IXA as a candidate for BBS. We exploited PCR-based techniques to clone a 8473-nt cDNA for myosin IXA. A 7644-bp open reading frame predicts a protein with all the hallmarks of class IX unconventional myosins. Human Northern blot analysis and in situ hybridization of mouse embryos reveal that myosin IXA is expressed in many tissues consistent with BBS. Intron/exon boundaries were identified, and myosin IXA DNA and RNA from BBS4 patients were evaluated for mutation.

Opposite orientations of an inverted duplication and allelic variation at the mouse agouti locus.

The mouse agouti protein is a paracrine signaling molecule that causes yellow pigment synthesis. A pale ventral coloration distinguishes the light-bellied agouti (AW) from the agouti (A) allele, and is caused by expression of ventral-specific mRNA isoforms with a unique 5' untranslated exon. Molecular cloning demonstrates this ventral-specific exon lies within a 3.1-kb element that is duplicated in the opposite orientation 15-kb upstream to produce an interrupted palindrome and that similarity between the duplicated elements has been maintained by gene conversion. Orientation of the palindrome is reversed in A compared to AW, which suggests that mutation from one allele to the other is caused by intrachromosomal homologous recombination mediated by sequences within the duplicated elements. Analysis of 15 inbred strains of laboratory and wild-derived mice with Southern hybridization probes and closely linked microsatellite markers suggests six haplotype groups: one typical for most strains that carry AW (129/SvJ, LP/J, CE/J, CAST/Ei), one typical for most strains that carry A (Balb/cJ, CBA/J, FVB/N, PERA/Rk, RBB/Dn); and four that are atypical (MOLC/Rk, MOLG/Dn, PERA/Ei, PERC/Ei, SPRET/Ei, RBA/Dn). Our results suggest a model for molecular evolution of the agouti locus in which homologous recombination can produce a reversible switch in allelic identity.

Neomorphic agouti mutations in obese yellow mice.

Several dominant mutations of the mouse agouti coat colour gene have pleiotropic effects that include obesity and a yellow coat. The Ay allele is caused by a large deletion that affects the expression of several contiguous genes. We show that three other obesity-associated agouti mutations, Aiy, Asy and Avy, are due to different molecular alterations that result in ubiquitous expression of a chimaeric RNA that encodes a normal agouti protein. The Aiy and Avy alleles are caused by insertion of an intracisternal A particle element 1 kb or 100 kb, respectively, upstream of agouti coding sequences. These results provide a model for other genes that show allele-specific imprinting, and demonstrate that molecular mechanisms typically responsible for activation of proto-oncogenes can also lead to other disease phenotypes.

Pleiotropic effects of the mouse lethal yellow (Ay) mutation explained by deletion of a maternally expressed gene and the simultaneous production of agouti fusion RNAs.

Heterozygosity for the mouse lethal yellow (Ay) mutation leads to obesity, increased tumor susceptibility and increased activity of the agouti coat color gene; homozygosity for Ay results in embryonic death around the time of implantation. Although these pleiotropic effects have not been separated by recombination, previous studies have suggested that the dominant and recessive effects result from distinct genetic lesions. Here we use a combination of genomic and cDNA cloning experiments to demonstrate that the Ay mutation is caused by a 120 kb deletion which lies centromere-proximal to the agouti coat color gene. The deletion removes coding but not 5' untranslated sequences for a ubiquitously expressed gene predicted to encode a protein similar in sequence to an RNA-binding protein, which we named Merc, for maternally expressed hnRNP C-related gene, but have renamed Raly, since the gene is nearly identical to one reported recently by Michaud et al. (Gene Dev. 7, 1203-1213, 1993). The Ay deletion results in the splicing of Merc/Raly 5' untranslated sequences to agouti protein-coding sequences, which suggests that ectopic expression of the normal agouti protein by the Ay fusion RNA is responsible for the pleiotropic effects associated with heterozygosity for Ay. We find that Merc/Raly RNA is present in the unfertilized egg and is also transcribed in preimplantation embryos. Using a PCR-based assay to determine the genotype of individual embryos from an Ay/a x Ay/a intercross, we show that, in the absence of zygotic Merc/Raly expression, Ay/Ay embryos develop to the blastocyst stage, but do not hatch from the zona pellucida or form trophoblastic outgrowths. Injection of a Merc/Raly antisense oligonucleotide into non-mutant embryos blocks development prior to the blastocyst stage, and can be rescued by coinjection of a Merc/Raly transgene. These results suggest that maternal expression of Merc/Raly plays an important role in preimplantation development and that its deletion of is sufficient to explain Ay-associated embryonic lethality.

Differences in dorsal and ventral pigmentation result from regional expression of the mouse agouti gene.

The agouti coat color gene encodes a paracrine signaling molecule that controls the production of yellow and black pigment by melanocytes within hair follicles. Some agouti alleles affect the dorsum and ventrum independently, which has provided the basis for speculation that agouti gene action in different regions of the body is controlled by distinct genetic loci that are closely linked. Using a combination of cDNA cloning and RNA expression studies, we find that alternative isoforms of agouti mRNA contain different noncoding first exons located 100 kb apart, whose patterns of expression indicate independent control by regulatory elements that are either ventral specific or hair cycle specific. These results demonstrate that the apparent genetic complexity of the agouti locus is explained by the existence of multiple regulatory elements exerting control over a single coding sequence and provide a conceptual basis for understanding differences in dorsal and ventral hair coloration in many mammalian species. The ventral-specific agouti isoform represents an example of a transcript whose expression is restricted to ventral skin and provide an approach to investigate the mechanisms by which dorsal-ventral differences in gene expression are established and maintained.

The mouse lethal nonagouti (a(x)) mutation deletes the S-adenosylhomocysteine hydrolase (Ahcy) gene.

The lethal nonagouti (a(x)) mutation is a hypomorphic allele of the agouti coat color locus which, when homozygous, also leads to embryonic death around the time of implantation. To understand the molecular basis of these phenotypes, we identified and cloned a deletion breakpoint junction present in the ax chromosome. Long range restriction mapping demonstrated a simple deletion of approximately 100 kb, which does not affect agouti coding sequences, but begins only 4 kb 3' of the last exon, and thus may affect coat color by removing an agouti 3' enhancer. The Ahcy gene, which codes for the enzyme S-adenosylhomocysteine hydrolase (SAHase), is contained within a 20 kb region within the a(x) deletion. SAHase RNA and protein were detectable in early blastocysts and in embryonic stem cells, respectively, and analysis of embryos derived from an a(x)/a x a(x)/a embryo intercross indicated that a(x)/a embryos die between the late blastocyst and early implantation stages. Treatment of cultured embryos with an SAHase inhibitor, 3-deazaaristeromycin, or with metabolites that can result in elevated levels of cellular SAH, resulted in an inhibition of inner cell mass development, suggesting that loss of SAHase activity in a(x)/a(x) embryos is sufficient to explain their death around the time of implantation.

Cloning of the mouse agouti gene predicts a secreted protein ubiquitously expressed in mice carrying the lethal yellow mutation.

The mouse agouti gene controls the deposition of yellow and black pigment in developing hairs. Several dominant alleles, including lethal yellow (Ay), result in the exclusive production of yellow pigment and have pleiotropic effects that include obesity and increased tumor susceptibility. In an interspecific backcross, we established genetic limits for the agouti gene and found that the Ay and the lethal non-agouti (ax) allele were not separated from a previously identified probe at the breakpoint of the Is1GsO chromosomal rearrangement. Using the Is1GsO probe, we isolated the agouti gene, and find that it has the potential to code for a secreted protein expressed in hair follicles and the epidermis, and that the level of expression correlates with the synthesis of yellow pigment. In the Ay mutation, there is a chromosomal rearrangement that results in the production of a chimeric RNA expressed in nearly every tissue of the body. The 5' portion of this chimeric RNA contains highly expressed novel 5' sequences, but the 3' portion retains the protein-coding potential of the nonmutant allele. We speculate that dominant pleiotropic effects of Ay are caused by ectopic activation of a signaling pathway similar to that used during normal hair growth.

Protein export from the mitochondrial matrix.

Assembly of a functional mitochondrion requires import of proteins from the cytosol and export of proteins from the matrix. Most previous studies have focused on the import pathway followed by nucleus-encoded proteins. However, it is now clear that proteins encoded in the nucleus as well as those encoded in the mitochondrion also move from the matrix into and across the inner membrane, a process defined here as export. These exported proteins are found in at least three cellular locations: the inner mitochondrial membrane, the intermembrane space and the cell surface. Here, we consider the pathways for export and the relationships between import and export.

The beta-adrenoceptor-coupled adenylate cyclase system in rat C6 glioma cells. Deamplification by isoproterenol and oxaprotiline.

The beta-adrenoceptor-coupled adenylate cyclase system in rat C6 glioma cells displays many characteristics observed in brain tissue: using nonlinear regression analysis of agonist competition binding curves, we demonstrated that the bulk of beta-adrenoceptors show high nanomolar affinity for isoproterenol; like in brain tissue, Gpp(NH)p does not shift agonist competition binding curves to the right; and the agonist isoproterenol rapidly downregulates the number of beta-adrenoceptors and deamplifies the norepinephrine signal. However, unlike in brain tissue, where (-)-oxaprotiline fails to decrease the number of beta-adrenoceptors and to desensitize the cyclic adenosine monophosphate generating system, it desensitizes the beta-adrenoceptor-coupled adenylate cyclase system in C6 glioma cells. Determination of the relative steady-state levels of beta-adrenoceptor messenger ribonucleic acid (mRNA) by Northern blot analysis showed a twofold increase in the steady-state levels of the mRNA at 30 minutes following exposure to (-)-isoproterenol or (-)-oxaprotiline. At 48 hours, basal values of mRNA were observed at a time when beta-adrenoceptors were maximally decreased. Further experiments on transcriptional activation, and mRNA stability and translation will be required to unravel the complexity of agonist-dependent and agonist-independent regulation of beta-adrenoceptor density and function.

Ethidium bromide fluorescence of 28S ribosomal RNA can be used to normalize samples in northern or dot blots when analyzing small drug-induced changes in specific mRNA.

Quantitative analysis of Northern blots is frequently accomplished with the aid of an internal standard. Most common is probing for an additional message the steady-state levels of which do not change in response to the experimental conditions and the signal of which is sufficiently removed from that of the target gene after gel electrophoresis. However, this strategy is not always feasible. When total RNA is immobilized on nylon, 28S ribosomal RNA on the blot can be used as an internal standard and quantitated by scanning the negative photograph of the blotted RNA stained with ethidium bromide. This procedure can also be used for RNA dot blots. The method is quick, reliable, will work with laser or white-light densitometers, and can serve as a universal internal standard, eliminating the need for additional probes.

Dose-dependent down-regulation of beta-adrenoceptors by isoproterenol in rat C6 glioma cells.

Nano- and micromolar isoproterenol concentrations were compared by studying cyclic AMP, beta-adrenoceptor density and beta 1-adrenoceptor mRNA in rat C6 glioma cells. 1 microM isoproterenol significantly changed all parameters at 15-30 min. The beta 1-antagonist metoprolol attenuated the response. No effects of nanomolar isoproterenol on these early changes were observed, although the density of beta-adrenoceptors was significantly reduced beginning at 12 h. The results indicate a different process for beta-adrenoceptor desensitization in C6 cells following physiologically low agonist concentrations.

Mitochondrial import of cytochrome c oxidase subunit VIIa in Saccharomyces cerevisiae. Identification of sequences required for mitochondrial localization in vivo.

Subunit VIIa of yeast cytochrome c oxidase is a small (59 amino acids) protein of the inner mitochondrial membrane that lacks a cleavable amino-terminal presequence. To identify regions within this polypeptide that are essential for its import, gene fusions were constructed using a leader peptide substitution vector (pLPS) developed in this laboratory (Glaser, S. M., Trueblood, C. E., Dircks, L. K., Poyton, R. O., and Cumsky, M. G. (1988) J. Cell. Biochem. 36, 275-287). In this vector, oligonucleotide sequences encoding all or part of subunit VIIa were fused in-frame with the coding region of mature cytochrome c oxidase subunit Va. The plasmid pLPS is ideal for assaying protein sequences for their ability to direct mitochondrial import in vivo since subunit Va's leader peptide is essential for import and because subunit V is required for cytochrome c oxidase activity and respiration. Strains containing these fusions but lacking both subunit V genes (COX5a and COX5b) were analyzed to determine whether the chimeric protein is directed to mitochondria. Our findings indicate that the amino-terminal 17 amino acids of subunit VIIa are sufficient to localize subunit Va to the mitochondrion and that a 6-amino acid-long region within the amino terminus (Gly8-Arg13) is essential. In addition, some import (approximately 10% of wild type) is observed with the highly charged carboxyl terminus of subunit VIIa, suggesting that the subunit may contain redundancy in its import information.

Intranuclear distribution of the human myeloid cell nuclear differentiation antigen in HL-60 cells.

Based on solubility properties, the human myeloid cell nuclear differentiation antigen exists as at least two distinct populations. Most is easily extracted from isolated nuclei in 0.35 M NaCl, while 20 percent resists such treatment. Compared to undigested nuclei, both the amount of myeloid cell nuclear differentiation antigen (MNDA) released from nuclei after DNase I treatment and the amount resisting further extraction in 0.35 M NaCl increased after DNA was digested with DNase I. Under these conditions, there was a concomitant decrease in the amount of MNDA that was extractable with 0.35 M NaCl. Mixing nuclear protein extracts that contain MNDA with nuclei from cells that do not express this protein demonstrated that the MNDA redistributes from the freely soluble form to the nuclear residual fraction as a consequence of DNase I digestion. These data are consistent with a model in which the amount of MNDA that is tightly bound to salt-washed nuclei is held constant in the presence of an excess of unassociated MNDA in the nucleus, and that the level of MNDA binding to this nuclear fraction increases in proportion to the extent of DNA damage resulting from DNase I digestion.

Poly(ADP-ribosyl)ation studies in situ.

We have studied the poly(ADP-ribosyl)ation of nuclear proteins in situ by examining the incorporation of [3H]NAD-derived ADP-ribose into polymers. We have devised a way to deliver [3H]NAD to cells growing in vitro, and we have determined the kinetics of uptake and incorporation into nuclear proteins using this delivery system. Incorporation into the histone fraction, known acceptors of poly(ADP-ribose), was examined and shown to be sensitive to the poly(ADP-ribose) polymerase inhibitor 3-aminobenzamide. Polyacrylamide gel electrophoresis of 3H-labeled proteins revealed radioactivity associated with known poly(ADP-ribose)-accepting proteins such as poly(ADP-ribose) polymerase and histones. These results were confirmed when we immunoreacted gel-separated proteins with anti-(ADP-ribose) generated in our laboratory.

Two human colon tumor cell lines with similar nuclease sensitivities have different ethidium bromide binding characteristics.

Chromatin from two human colon adenocarcinoma cell lines (HT-29 and LoVo) showed similar digestion kinetics when sensitivities to DNase I and micrococcal nuclease were examined. Chromatin conformations were probed by examining the binding of ethidium bromide. A Scatchard plot revealed that both chromatins bound the same amount of ethidium bromide per mole of DNA, but the DNA from LoVo cells was more accessible to the intercalator. The results indicate that differences in chromatin conformation are not necessarily accompanied by different nuclease sensitivities.

Rapid method for preparation of nuclei from a mucin-producing solid tumor.

Nuclei were prepared from the human colon mucin-producing tumor GW-39. The method described involves incubation in beta-mercaptoethanol under mild conditions and results in the preparation of nuclei free of the mucin web that initially enmeshes them. A sensitive colorimetric immunoassay was used to confirm that nuclei prepared in this fashion were generally free of contaminating mucin. This method is mild, rapid, and should be applicable to other mucin-producing tumors.

A nuclear matrix antigen HeLa and other human malignant cells.

Antisera were obtained in rabbits to preparations of dehistonized chromatin from HeLa cells. By complement fixation assays, the antisera reacted with HeLa cell chromatin but only marginally with human placenta chromatin. The complement-fixing reactivity of the antisera was inversely related to the amount of dehistonized chromatin used for immunization. Immunochemical staining of electrophoretically separated chromosomal proteins transferred to nitrocellulose sheets revealed numerous antigens in chromatin preparations from several human tumors, placenta, and normal kidney. While immunoabsorption of the antisera with placenta chromatin removed some of the immunochemical staining, many of the electrophoretically separated antigens resisted repeated immunoabsorptions. However, further comparisons revealed that only one major protein antigen (band at an approximate molecular weight of 81,000) was represented in all the assayed human tumors while being absent from human placenta or kidney. Fractionation of HeLa cells into three cytoplasmic and several nuclear fractions showed that almost all the antigens recognized by antisera to dehistonized chromatin were nuclear. The antigenic protein with an approximately molecular weight of 81,000 was found associated with the nuclear matrix fraction.