1. Some experiences from the past
So far, almost 650 clones have been requested from the GGTC by interested scientists. However, not everyone is used
to analyzing gene trap mice. Therefore some of the molecular analyses went quite fast, whereas others were not successful
even after several months. In order to prevent you from making the same mistakes that have been made by others
before, we are trying to summarize all the things that we are aware of, which had happened in the past during the
insertion event of gene trap vectors in ES cells. In particular, we will emphasize the differences between
plasmid DNA and proviruses. Both have their advantages and disadvantages and naturally, these will determine the molecular
analysis of choice for a given integration.
1.1 Plasmid vs. Retrovirus
More than 5000 of our clones have been produced by using the plasmid vectors pT1ßgeo or pT1ATGßgeo.
These vectors are still most efficient in terms of trapping new genes and most probably integrate into the genome by the cellular DNA repair machinery.
Retroviruses are easier to handle and will usually produce single-copy insertions, however they tend to reach
saturation more quickly due to micro-homologies to the LTRs which are quite frequent in the mouse
genome. The fact that plasmid DNA tends to integrate in a given locus in concatemers often leads to difficulties
even during RACE PCR. In case multiple expressed genes were hit and RACE products have similar sizes, direct
sequencing of RACE products may be impossible, due to overlaying signals. Therefore, many of those are eliminated
already at this stage and will usually not appear in the GGTC database. Tandem or multiple insertions will compete for internal
primers during RACE or genomic PCR. A second feature, that makes the analysis of
plasmid vectors unpredictable is the so-called "chew-back" by exonucleases from both sides of the linearized
vector during the course of integration. We have observed the loss of almost all the 1688 bp of the en2 intron (clone
W027B02; Rantanen et al., 2002) as well as the complete retainment of the en2 intron (clone W052E02; unpubl.). Small
rearrangements within the en2 intron were observed (clone W036C08; Hitz et al., 2005) as well as major
rearrangements of the whole genomic locus (clone W035F03; submitted). Basically, what is known of a
plasmid-generated mutation before starting the molecular analysis is that neomycin is functional and therefore
complete. Additionally the lacZ cassette should also be complete, since the first 5'RACE-primer matches here. Even
the en2 exon 2 (between 1688-1875 bp in pT1ßgeo) is most likely intact, as the nested 5'RACE primer
is complementary to it (sequence information available). However, lacZ may still be dysfunctional, as a result of vector methylation, en2 splice site loss and cryptic splicing
events – the latter were often found to be exon-integrations. In general, lacZ staining can be detected only in
30% of trapped ES cells (please note that some of our vectors do not contain any reporter or a GFP instead of lacZ).
Chew back unfortunately leads to the failure of inverse- or linker-mediated PCR methods, which rely on the presence of
a particular restriction site. In one particular case of a plasmid-generated trap, a genomic locus was found to be
completely rearranged in a way that exons which are located 3' in the gene were found closer to the
5' end after vector integration (clone W035F03; in preparation). In summary, the molecular analysis of a
given mutation is easier if a retrovirus was used, since a defined situation with a single-copy
integration and the retainment of the entire provirus can be found.
2. Clone-confirmation: What we do, what you do
Generally, all clones that we have sent out since January 2004 have been sequence-confirmed by
RACE-PCR - retroviral traps also by splinkerette PCR. Please ask for
your particular clone, if we have not sent you any information yet. However, sequence-confirmation might take
some time. We can not guarantee that there has not been any mix-up of cells or sequences on our side. Below we listed the possibilities you
have, starting with the fastest way, which is RT-PCR.
The primers you may use for a given vector can be the same that we use for 5'RACE PCR.
Currently, 40 different vectors are in use.A list of our specific RACE PCR primers can be found here (RACE protocol).
If you are in doubt, please check our website, which vector has been utilized in similar cases.
The following primers are recommended for genomic PCR:
- lacZ: CAAGGCGATTAAGTTGGGTAACG
- U3Ceorev2: agggttgtgttgatacaagtccagga
3. The possibilities you have
3.1 RT PCR
All the genes that we trap are expressed in ES cells, therefore you should be able to detect a
fusion transcript in total RNA made from the clone we sent to you. We recommend random-primed cDNA, since splice-acceptor traps tend to locate 5' in a cDNA.
RACE protocols can be found here (RACE protocol). Please note that
we do not use any kits, since this would be too expensive for large scale.
3.3 Genomic PCR: introns can be huge!
In some cases, vectors had integrated in 100 kb introns (clone W008G09, unpubl.). Unless there
is a splinkerette PCR result available, it is not possible to tell, where exactly an intron integration had
occured. In case of alternative splicing in ES cells, the integration could have happened even in a more
downstream intron. In these cases, we advise to move to step 3.6. (RFLP). If your intron is not over
15-20 kb, we recommend long-range genomic PCR. First, design good primers with a temperature of at least 68°C for
each 3-5 kb in the “intron-of-integration” and PCR using vector-specific primers at different
extension times – we urgently suggest the primer 3 workpackage which is freely available and allows masking out repetitive sequences. We use a two-step PCR with Advantage Taq and subsequently
clone the product in a TOPO vector for sequencing. Make sure you include wildtype mouse DNA as a negative control
(not provided by us).
3.4 Inverse PCR (retroviruses only!)
A detailed protocol for inverse PCR can be found here (inverse PCR). Inverse
PCR cannot be applied to plasmid vectors because of "chew-back", but it works well for retroviruses and returns intron sequence external to the integration.
3.5 Splinkerette PCR (some retroviruses only!)
Splinkerette PCR involves digestion with a particular restriction enzyme and ligation of an oligo, which folds back intramolecularly to serve
as a primer for the PCR reaction in the third step. This requires retroviral vectors and works only for some so far.(splinkerette PCR).
If all methods fail (besides RT-PCR and 5’ RACE of course!) you should think of generating external probes for southern blotting. These
can be easily made by genomic PCR – just make sure to avoid having repetitive sequences in your probe (http://www.repeatmasker.org/). In this way, you will
cover large genomic regions depending only on the restriction enzymes you use. This will give you a good indication on where your vector is located with respect
to certain restriction sites. Find a restriction map for each vector here (restriction map). Keep in mind
that there are numerous polymorphisms between C57BL/6 and 129 mouse strains, which can be misleading in southern plots. Make sure to include DNA from both pure 129 and C57 in your experiment.
3.7 Phage and other libraries
If a genomic locus could not be analyzed by the methods above, a phage library from a given ES clone should be established and a screen with probes internal to the vector done. Follow the usual protocols ( see 3.7.1).
3.7.1 Colony hybridization
If you know the approximate size of your genomic fragment from a southern blot with internal probes, cut with this specific enzyme, and subclone it in any plasmid to carry out colony-hybridization with the same internal probe. ( Short Protocols in Molecular Biology (1994): Frederick M. Ausubel , Roger Brent & Robert E. Kingston (eds.) Wiley ; Molecular cloning. A Laboratory Manual. Sambrook & Russell 3rd Edition. 2001. Cold Spring Harbor Laboratory Press )
3.8 Multiplex PCR/TaqMan assay
Once the boundary between the gene trap vector and the intron it is located in has been sequenced, it is advisable and straightforward to
establish a genotyping assay which produces smaller bands in the range of 150-300bp and a second band for the wildtype locus, which will reliably distinguish
between wildtype and trapped locus. Especially if you have to analyze early embryonic phenotypes, a robust PCR assay is inevitable due to the small amounts of
tissue obtained. For single-copy integration (TaqMan Assay), we have established a quantitative PCR assay with the plasmid vectors pT1ßgeo and pT1ATGßgeo. This assay distinguishes one or two copies of betageo in respect to a single-copy reference gene in the mouse genome (Floss et al., 2002).