Tim Thomas

Efficiency assessment of the gene trap approach.

The trapping of genes in murine embryonic stem (ES) cells offers three features in one experimental approach: 1) analysis of the expression patterns of unknown genes by using a simple staining method, 2) rapid cloning of unknown genes, and 3) generation of mutant mouse lines. We performed a gene trap screen aimed at the discovery of new genes regulating embryonic development. We have processed 209 gene trap events for expression patterns in chimeric murine embryos. Randomly tested, β‐galactosidase‐positive ES cell clones resulted in vivo in 35% gene trap events showing no β‐galactosidase activity, 39% gene trap events with ubiquitous β‐galactosidase activity, and 26% gene trap events showing β‐galactosidase activity restricted to specific cell types or organs. In vitro preselection reduced gene trap events with ubiquitous β‐galactosidase activity to 10% and increased the gene trap events with restricted β‐galactosidase activity to 64%, making the screening procedure for genes expressed in a restricted manner 2.5‐fold more efficient. In five of the seven gene trap insertions into genes in which the expression pattern during embryogenesis was known, the β‐galactosidase marker gene reproduced faithfully the expression pattern of the trapped gene. 5′‐Rapid amplification of cDNA ends (5′‐RACE) of 28 gene trap events revealed 19 novel mouse genes, 8 known mouse genes, and 1 random trans‐splicing event. Twelve of the 25 mouse lines that crossed to homozygosity showed overt abnormalities. The genomic structure was investigated in four of these gene trap events, which caused obvious abnormalities. In all four cases, the splice‐acceptor gene trap construct was inserted into an exon. One of the 13 gene trap events that did not result in overt abnormalities was examined for the presence of wild‐type mRNA. Homozygous animals were found to produce normal levels of wild‐type mRNA. Evidently, gene trapping does not always provide all three of the features mentioned above. In this paper, we discuss the efficiency of gene trapping and ways in which some problems may be overcome. Dev. Dyn. 1998;212:171–180.

Compensation for a gene trap mutation in the murine microtubule‐associated protein 4 locus by alternative polyadenylation and alternative splicing

One of the features expected of the gene trap approach is the functional mutation of a gene, allowing its loss‐of‐function phenotype analysis. We have mutated the murine microtubule‐associated protein 4 (MAP‐4) locus by inserting a splice‐acceptor gene trap construct. Because the MAP‐4 gene has been cloned, sufficient information is available to investigate the efficiency of the gene trap insertion in disrupting the protein‐coding region. The fusion mRNA contains the first 905 bases of the MAP‐4 mRNA and is expected to code for a truncated, nonfunctional MAP‐4 protein missing, among others, the microtubule‐binding domain. Activity of the lacZ marker gene of the gene trap construct was observed in all tissues throughout development and in all cells examined in adult animals. However, some cells and tissues showed higher levels of activity than others: for example, blood vessel endothelium, heart, aspects of the developing nervous system, surface ectoderm of embryonic day 11.5 embryos, and the ependymal layer and blood vessel endothelium in adult brain. MAP‐4 binds to microtubules and is thought to modulate their stability. It is expressed differentially in different tissues as 5.5‐kb, 6.5‐kb, 8‐kb, 9‐kb, and 10‐kb mRNA species from a single copy gene in mice. Northern hybridization with a 5′, MAP‐4‐specific probe revealed a 3.3‐kb splice variant, which has not been described previously, that was expressed as the most abundant MAP‐4 mRNA species in the brain but not in other tissues. Mice homozygous for the reported gene trap insertion in the MAP‐4 locus (MAP‐4gt/gt) are viable and appear to be phenotypically normal. They exhibited normal levels of all MAP‐4 mRNA species in brain and kidney, showing that the simian virus 40‐polyadenylation signal of the gene trap construct was ignored and also showing compensation for the gene trap insertion by splicing around the gene trap construct. Dev. Dyn. 1998;212: 258–266.

Distribution of a murine protein tyrosine phosphatase BL-β-galactosidase fusion protein suggests a role in neurite outgrowth.

We have generated a gene trap insertion into the protein tyrosine phosphatase‐BL (PTP‐BL) locus, which produces a fusion of the N‐terminal half of PTP‐BL with β‐galactosidase. During development, β‐galactosidase activity was seen in all epithelial cells: strong staining was observed in the stomach and kidney epithelium, the ependymal layer of the central nervous system, and the surface ectoderm. Particularly prominent β‐galactosidase activity was seen in the peripheral nervous system, which correlated with neurite outgrowth. In epithelial cells, staining was seen in the apical portion of the cells. In nerves, β‐galactosidase activity was associated with growth cones as well as with Schwann cells. This suggests that the amino‐terminal portion of PTP‐BL contains sequences sufficient to target the fusion protein to specific subcellular compartments. In situ hybridization with a PTP‐BL probe demonstrated that all tissues in which β‐galactosidase activity was seen were genuine sites of expression of the PTP‐BL gene, although differences in the stability of the PTP‐BL protein and the PTP‐BL‐β‐galactosidase fusion protein may exist. The distribution of β‐galactosidase activity in the peripheral nervous system, together with the structure of the wild‐type protein, suggests that this phosphatase may have a role in regulation of the cytoskeleton during the development of the peripheral nervous system. Dev. Dyn. 1998;212:250–257.

A new gene trap construct enriching for insertion events near the 5' end of genes.

The gene trap approach is based on the integration of a gene trap vector into the genome. This can be done either by electroporation of a plasmid construct or by infection with a viral vector. Commonly used viral gene trap vectors have been shown to select for integrations near the 5′ end of genes. To date, no plasmid vector with a similar tendency has been reported. In this paper we describe a new plasmid vector, pKC199βgeo. This vector contained a short splice acceptor fragment from the Hoxc9 gene, a full length lacZ gene, including an ATG, and a reduced activity, mutant neomycin phosphotransferase gene as a selectable marker. This vector enriched the population of trapped genes in our gene trap screen for insertion events in the 5′ end of genes. In the two cases examined the β-galactosidase activity pattern accurately reflected the endogenous promotor activity.