We examined the utility of a SB transposon-base gene trap vector that expresses the Tet-off transcriptional transactivator in the patterns of trapped genes in vitro and in vivo. The gene-trap tTA demonstrated full utility in cultured cells, having the ability to insert in genes, trap the splicing of those genes to express the tTA transcription factor which, in turn, could activate a TRE-regulated transgene in a doxycycline-dependent manner. Likewise, analysis of gene-trap tTA transposon insertions into mouse genes revealed that the vector system can completely disrupt the expression of a mouse gene in some cases and can tissue-specifically activate expression of a TRE promoter-regulated transgene in a doxycycline-dependent manner in vivo. While ultimately this vector system can function as designed, several observations suggest we need a more optimal approach for creating a resource of mice that express tTA in developmentally-controlled patterns.
Germline SB transposon mobilization rates in the mouse germline are variable and ranged from less than one in twenty to nearly three events per gamete. An evaluation of several different transposon-transgenic lines has given us insight into the potential the SB system to perform regional mutagenesis throughout the genome. The genomic location of the transposon donor site, along with copy number, impacts the germline mobilization rate from any concatemer (see Table ). This suggests that not all parts of the genome are equally accessible to SB transposition. Understanding the mechanisms behind this "mobilization position effect" will likely require an increased understanding about molecular regulators of SB
transposition, but may include a well known director of position effect, heterochromatinization [30
As demonstrated in previous studies and here, SB
transposon-based gene trap vectors can function in vivo
to disrupt the expression of mouse genes. The T2/GT2/tTA vector can completely ablate the expression of a mouse gene as demonstrated by the null mutation that was generated in the mouse Car12
gene. Not all gene insertions completely disrupt gene expression however, as in the case of insertion 02A-0002 into ENSMUSG00000066992
. While hypomorphic mutations may be valuable to discover the function of an essential mouse gene, this has lead to ideas on how to improve the efficacy of future vectors. Splice acceptors derived from different genes, termination sequences (polyadenylation signals) and other functional sequences have been used in several ES cell plasmid- and retroviral-based gene-trapping studies with varying efficiencies [31
]. Drawing data from other groups will lead to improved vectors for trapping mouse genes and perhaps lead to vectors to trap certain classes of genes [34
The activation of luciferase by the tetracycline-controlled transactivator varied rougly 1000-fold (Fig. ) when expressed from the IRES in the gene-trap tTA. Once again, improvements to the functional components of the gene trap vector may lead to more reliable and greater expression of the transactivator. Alternative IRES elements like the 9-nucleotide IRES element from the murine Gtx
homeodomain gene [35
], or eliminating the IRES entirely, may allow for more consistent tTA expression.
The brilliant pattern of in vivo luciferase expression seen if Fig. was initially thought to originate from insertion in the MacF1 gene. Published reports suggested that neither the bone marrow, nor hematopoietic cells are in the normal expression domain of this gene. Upon further molecular characterization, we discovered that this gene was not spliced into the gene-trap tTA, and by Southern blot, found it was tightly linked to other insertions as well as the concatemer. It seems more likely that the pattern of tTA-dependent luciferase activation seen in carriers of the 03A-0241 insertion comes from one of these linked transposons. Efforts to find the source of this expression by 5'RACE were unfortunately unsuccessful. We propose that in the case of insertion 03A-0241, the concatemer translocated to chromosome 4 first, then a single copy of T2/GT3/tTA hopped locally into MacF1. We have observed similar and other types of genomic rearrangements associated with transposition from the chromosome-11 donor site in other mice (A.M.G., manuscript in preparation). These observations would suggest that gene-trap tTA mobilization in the germline may not be the most efficient means of generating a resource of mice with tissue-specific tTA expression patterns. It is not clear, however, that these large donor site insertions are not rare events that could be tolerated. Independent insertions, unlinked to the concatemer or other transposons, are frequently recovered. By pre-selecting mice that do not inherit the concatemer by Southern blot or PCR analysis, a screen would largely avoid the complications of linked transposons.
An alternative and relatively facile approach would be to inject a linearized plasmid harboring the gene-trap tTA transposon, along with in vitro
-transcribed transposase messenger RNA, into one-cell mouse embryos. In the one-cell embryo, the transposase mRNA is translated and transposase catalyzes integration into the mouse genome at rates that are two to three-fold higher than standard pronuclear injection of naked DNA [3
]. This method produces transgenic mice with insertions distributed randomly across the genome, and multiple linkage groups are frequently obtained from a single injection [3
]. This would be another method to avoid the complications of linked insertions and concatemers.
Finally, using the GAL4
/UAS enhancer trap systems used in the fly, expression-based screening has been useful for identifying genes with similar or overlapping expression patterns [36
] and for tissue-specific overexpression studies [37
]. As we have shown here, a transgenic mouse expressing firefly luciferase under the control of the tet-response enhancer/promoter (TRE) can be coupled with the gene-trap tTA system to identify in vivo
patterns with an intensified CCD camera system. In a screen, this type of imaging could be used to identify gene trap insertions that lead to tTA expression in a particular tissue or developmental stage, even in utero
]. Alternatively, mouse strains for other TRE-regulated transgenes have been created, allowing for tTA-dependent expression of β-galactosidase [40
] or GFP [41
]. If the efficiency of the gene-trap tTA transposon can be improved, these strains could be used by any lab to identify tissue-specific patterns.