Faculty Bibliography

More than 10 megabases of contiguous genome sequence have been submitted to the databases by the Caenorhabditis elegans Genome Sequencing Consortium. To characterize the genes predicted from the sequence, we have developed high resolution FISH for visualization of mRNA distributions in whole animals. The high resolution and sensitivity afforded by the use of directly fluorescently labelled probes and confocal imaging permitted mRNA distributions to be recorded at the cellular and subcellular level. Expression patterns were obtained for 8 out of 10 genes in an initial test set of predicted gene sequences, indicating that FISH is an effective means of characterizing predicted genes in C. elegans.

A strain of Caenorhabditis elegans was constructed that permits selection of dominant or sex-linked mutations that transform XO animals (normally male) into fertile females, using a feminizing mutation, tra-2(e2046gf), which by itself does not sexually transform XO males. Twenty-three mutations were isolated after chemical mutagenesis and found to fall into both expected classes (four dominant tra-1 mutations and eight recessive xol-1 mutations) and novel classes. The novel mutations include 10 second-site mutations of tra-2, which are called eg mutations, for enhanced gain-of-function. The tra-2(gf, eg) alleles lead to complete dominant transformation of XO animals from fertile male into fertile female. Also isolated was a duplication of the left end of the X chromosome, eDp26, which has dominant XO lethal and feminizing properties, unlike all previously isolated duplications of the X chromosome. The properties of eDp26 indicate that it carries copies of one or more numerator elements, which act as part of the primary sex-determination signal, the X:A ratio. The eDp26 duplication is attached to the left tip of the X chromosome in inverted orientation and consequently can be used to generate unstable attached-X chromosomes.

Sex in Caenorhabditis elegans (XX hermaphrodite, XO male) is determined by the X:A ratio, which is the ratio of X chromosome number to autosomal set number. Recent genetic results with X chromosome duplications have suggested that there may be only a small number of major numerator sites on the X chromosome that contribute to this ratio. Mapping of duplication endpoints delimited a region of less than 300 kb, likely to contain one such element. Cosmid clones from this region were tested for numerator activity by constructing transgenic lines carrying extra copies of each tested cosmid. Most cosmid arrays have no effect on the viability of either XX or XO animals. One cosmid array was found to be viable in XX animals, but lethal and feminizing in XO animals, consistent with it containing a major numerator element. Further experiments defined a region of 12-30 kb with apparent numerator activity, which is designated fox-1, 'Feminizing locus On X'. A cDNA clone hybridizing across part of this region encodes a predicted RNA-binding protein.

dpy-27 is an essential dosage compensation gene that acts to reduce expression of both hermaphrodite X chromosomes. The DPY-27 protein becomes specifically localized to the X chromosomes of wild-type XX embryos, but remains diffusely distributed throughout the nuclei of male (XO) embryos. In xol-1 mutant XO embryos that activate the XX mode of dosage compensation and die from inappropriately low X chromosome transcript levels, DPY-27 becomes localized to X. Therefore, sex specificity of the dosage compensation process is regulated at the step of DPY-27 X chromosome localization. DPY-27 exhibits striking similarity to proteins required for assembly and structural maintenance of Xenopus chromosomes in vitro and for segregation of yeast chromosomes in vivo. These findings suggest a link between global regulation of gene expression and higher order chromosome structure. We propose that DPY-27 implements dosage compensation by condensing the chromatin structure of X in a manner that causes general reduction of X chromosome expression.

A pseudogene related to the functional gene (FECH) for the heme biosynthetic enzyme ferrochelatase (ferroheme-protolyase; EC 4.99.1.1.) was isolated from a human genomic library using a ferrochelatase cDNA hybridization probe. The pseudogene shows > 80% overall nucleotide sequence identity to the functional gene (including the 3' untranslated region and polyadenylation signals) but contains no intronic sequences in the region corresponding to the open reading frame of expressed ferrochelatase. Furthermore, the pseudogene sequence contains small deletions and insertions creating frameshifts and numerous termination codons, indicating that it does not encode a functional polypeptide. Northern blot analysis using pseudogene-specific probes failed to demonstrate transcripts in samples of human erythroid cell RNA in which ferrochelatase mRNA was readily detected. Southern blot experiments using restriction endonuclease-digested human genomic DNA probed either with ferrochelatase-specific cDNA fragments or pseudogene-specific genomic sequences confirmed the presence of distinct loci for the expressed and nonfunctional genes, respectively. Localization of the human ferrochelatase pseudogene to 3p22-p23 was determined by fluorescent metaphase chromosomal hybridization in situ using three genomic clones in lambda EMBL3 spanning a contiguous region of approximately 30 kb. This newly identified locus, distinct from the expressed FECH gene, on 18q22, is characteristic of a processed human pseudogene. The existence of the ferrochelatase pseudogene has practical implications for the molecular analysis of mutations responsible for erythropoietic protoporphyria in man.

In constructing complete human chromosome maps, the relative order of markers and their precise chromosomal location will be combined. Multicolour in situ hybridisation, in which two probes are simultaneously hybridised to chromosomes and subsequently distinguished, potentially will provide both types of information. Using this technique, we have produced an ordered map of eight human chromosome 3 DNA markers, using small, single-copy probes that can detect target sequences ranging in size from 4 kb to as little as 500 bp.

A scheme for rapidly mapping chromosome rearrangements relative to the physical map of Caenorhabditis elegans is described that is based on hybridization patterns of cloned DNA on meiotic nuclei, as visualized by fluorescent in situ hybridization. From the nearly complete physical map, DNA clones, in yeast artificial chromosomes (YACs), spanning the rearrangement breakpoint were selected. The purified YAC DNAs were first amplified by degenerate oligonucleotide-primed polymerase chain reaction, then reamplified to incorporate fluorescein dUTP or rhodamine dUTP. The site of hybridization was visualized directly (without the use of antibodies) on meiotic bivalents. This allows chromosome rearrangements to be mapped readily if the duplicated, deficient or translocated regions do not pair with a normal homologous region, because the site or sites of hybridization of the probe on meiotic prophase nuclei will be spatially distinct. The pattern, or number, of hybridization signals from probes from within, or adjacent to, the rearranged region of the genome can be predicted from the genetic constitution of the strain. Characterization of the physical extent of the genetically mapped rearrangements places genetic landmarks on the physical map, and so provides linkage between the two types of map.

The meiotic segregation of the holocentric chromosomes of Caenorhabditis elegans in both spermatogenesis and oogenesis is described. The extended kinetochore typical of the mitotic chromosome could not be differentiated on meiotic bivalents; instead microtubules appeared to project into the chromatin. The meiotic spindles formed during spermatogenesis contain centrioles and asters, while in oogenesis the spindles are acentriolar and barrel shaped. The formation of the acentriolar spindle was studied in fixed specimens by anti-tubulin immunofluorescence. Microtubule arrays were seen first to accumulate in the vicinity of the meiotic chromosomes prior to congression. At later stages, elongated spindle structures up to 13 mu in length were observed parallel to the surface of the embryo. Further development of the spindle appeared to involve its shortening into a barrel shape and rotation so that one spindle pole was opposed to the membrane. By anaphase the pole-to-pole spindle length reached a minimum of 3-4 mu. One end of each chromatid in the meiotic bivalent was labelled by in situ hybridization of a probe DNA to show that in oogenesis the chromatids were associated end-to-end in the bivalent. Furthermore, either the right or the left ends of the homologues could be held in association. At metaphase I the bivalents were oriented axially, such that kinetic activity was restricted to one end of each pair of sister chromatids. At metaphase II the chromosomes were also aligned axially.

Using fluorescence in situ hybridisation (FISH) the chromosomal location and relative order of six human chromosome 3 probes has been determined. The sensitivity of the technique has enabled the relative mapping of probes carrying inserts as small as 500 basepairs (bp), thus allowing the following proximal-distal probe order to be proposed: D3S30 (3p13-14), D3S4 (3p13-14), D3S2 (distal 3p14), D3S32 (3p21), D3S48E (3p21-23), and D3S11 (3p22-23). These data combined with the deletion mapping data of other researchers raise the possibility that the loss of more than one region of the short arm of chromosome 3 may be important in the development of small cell lung cancer.