Jane Brennan and William C Skarnes 1 Introduction

Gene trapping in mouse embryonic stem (ES) cells offers a method to create random developmental mutants with a direct route to cloning and defining the expression pattern of the disrupted gene (1). Gene trapping involves the use of reporter gene constructs that are activated following insertion into endogenous transcription units. A number of plasmid- and retroviral-based vectors have been developed, which differ in their requirements for reporter gene activation (reviewed in refs. 2 and 3). "Promoter trap" vectors simply consist of a promoterless reporter gene that is activated following insertions in exons of genes. In contrast, "gene trap" vectors contain a splice acceptor sequence upstream of a reporter and are activated following insertions into introns of genes. Both promoter and gene trap insertions create a fusion transcript from which a portion of the endogenous gene may be readily cloned (4,5). The pattern of reporter gene activity can be monitored in ES-cell derived chimeric embryos (6) or in transgenic embryos following germline transmission (7). With two plasmid-based vectors, reporter gene expression has been shown to reflect accurately that of the endogenous gene (5,8). Ultimately, the function of the trapped gene can be tested following germline transmission. Using this approach, a number of embryonic lethal mutations and visible adult pheno-types have been isolated (4,5,7,9-11).

Before initiating a screen, it is important to consider carefully the design of individual vectors, their efficiency in detecting gene trap events, and their potential biases (discussed in refs. 2 and 3). For example, although gene trap vectors are more efficient at detecting insertions in genes than promoter trap

From: Methods in Molecular Biology, Vol. 97: Molecular Embryology: Methods and Protocols Edited by: P. T. Sharpe and I. Mason © Humana Press Inc., Totowa, NJ

vectors, they will favor the detection of genes composed of large intronic regions, and are likely to miss genes possessing few or no introns. Gene trap vectors are also constrained by the reading frame imposed by the splice acceptor sequence. One solution has been to incorporate the splice acceptor derived from Moloney murine leukemia virus (MoMuLv) env gene that is capable of splicing in all three reading frames simultaneously (12). However, splice acceptors with this property may be weak, so they may fail to mutate the gene at the site of insertion effectively. Alternatively, an internal ribosome entry site (IRES) may be used to initiate translation of the reporter gene independent of the upstream open reading frame (13). Our own answer to this problem has been to construct separate vectors in each of the three reading frames. Given that individual vector designs each have their own inherent biases, we recommend using a combination of vectors to ensure the most representative sampling of the genome.

Perhaps the most useful vectors are those based on the Pgeo reporter system (7). Pgeo encodes a polypeptide fusion possessing both P-galactosidase and neomycin phosphotransferase (neo) activities, thereby providing direct drug selection for gene trap events and obviating the need to screen through a large background of nongene trap events. However, there is some debate over whether neo driven by its own promoter may be better suited to trap genes activated at later stages of development. In this regard, the sensitivity of the drug selection marker becomes an important consideration. It has been shown, for example, that the original Pgeo reporter contained a mutation in neo and thus tended to preselect for genes expressed at high levels (8). Correction of this mutation has enabled the isolation of genes expressed at low levels in ES cells that are activated on differentiation.

The isolation of so-called white colonies (cells expressing less than detectable levels of P-galactosidase [Pgal] activity) does not necessarily indicate that the trapped gene is expressed at a low level, but rather may reflect inactivation of the Pgal enzyme in the resulting fusion. One important class of genes that produce inactive Pgal fusion products are those that encode N-terminal signal sequences (8). To capture specifically this class of genes, which include secreted and membrane-spanning proteins, we modified the conventional gene trap design by adding a type II transmembrane domain upstream of Pgeo. With the secretory trap vector pGT1.8TM, only fusions that acquire an N-terminal signal sequence will produce an active Pgal fusion. One surprising result to emerge from this study was the isolation of two independent insertions in the same gene in a sample of six cell lines, suggesting that gene trapping may be far less random than originally anticipated. This finding further emphasizes the need to recognize and ultimately overcome inherent biases imposed by individual vector designs.

This chapter will focus on the use of Pgeo-based plasmid vectors in mouse ES cells to screen for developmentally regulated genes in the mouse. Methods for the maintenance and electroporation of a feeder-independent ES cell line are described. To identify Pgal-positive clones, we employ a screening protocol that simply involves staining one set of duplicate wells with X-gal. This procedure can also be used to screen for genes induced or repressed in response to specific growth factors or other inducers of ES cell differentiation. Several methods have been used to clone a portion of the endogenous gene associated with gene trap and promoter trap insertions. These include the construction of cDNA libraries (10), inverse PCR

(4), ligation-mediated PCR (11), and 5' rapid amplification of cDNA ends (RACE)

(5). This chapter, provides a detailed method for 5' RACE cloning, used routinely in the authors' laboratory, in which a number of improvements on previously published protocols have been added (3,14). The generation of ES cell-derived chimeras by blastocyst injection or morulae aggregation is used to monitor expression patterns in embryos and to transmit the insertions to the germline of mice. These methods are not covered in this chapter, since they have been described extensively elsewhere (15-18). Finally, we have included a rapid dot-blot method for genotyping transgenic mice carrying gene trap insertions. This method can be used to analyze as many as 400 tail biopsies in a single day, and can reliably distinguish between heterozygous and homozygous animals based on signal intensity.

Was this article helpful?

0 0

Post a comment