Guidelines

Design and Assembly of Targeting Vectors

Please note: The design of a gene targeting experiment is a scientific issue and is not considered part of the technical cost of performing a gene targeting experiment. Some experimental approaches may be conceptually or technically novel, or involve propriety reagents, and thus may necessitate collaboration.

1. There are many factors to consider when designing a gene targeting vector, some of which can greatly affect the probability of a successful outcome. The type of gene targeting vector also depends on the experimental goals. For instance, the considerations involved in designing a vector to generate a conditional allele are different from those for making a cassette acceptor or knock-in allele. The vector design is also affected by the nature of the gene and the type or structure of the protein it encodes. In addition, the location of restriction enzyme sites as well as presence of repetitive DNA sequences need to be taken into account.

   If you have never performed a gene targeting experiment and lack the experience and knowledge necessary to make these judgments we strongly recommend that you identify a collaborator to assist you with the steps that are described below. There are several laboratories within the Vanderbilt University Medical Center that have extensive experience in the design of gene targeting vectors. We are happy to make recommendations. If you do not think collaboration with anyone is necessary but still want some advice you may obtain consultation on fee-for-service basis on the design of gene targeting vectors from Drs. Magnuson or Labosky, and on specific BAC or Red/ET recombineering issues from Dr. Mortlock. There is some written material that can help such as ‘Gene Targeting, A Practical Approach’, by A.L. Joyner or ‘Manipulating the Mouse Embryo’ by A. Nagy. However, for the most part these texts focus on general issues and often ignore factors that may be critical for the success of a specific type of experiment.

   Please keep in mind that even with an ‘optimally designed’ targeting vector some genetic loci may be difficult or impossible to target. Thus, there is no guaranteed outcome even if hundreds of colonies are screened.

   It may not even be necessary to perform a gene targeting experiment to obtain a knock-out mouse. There are several repositories of ES cells that have specific retroviral insertions that may have created null mutations. For instance, the Texas Institute for Genomic Medicine (http://www.tigm.org) has large libraries of mutant C57Bl/6N and 129/SvEvBrd ES cells. These libraries, which together contain over 640,000 mutant cell lines, represent over 13,000 genes. Also, the NIH funded ‘Knock-Out Mouse Project (KOMP) is making rapid progress so the KOMP repository (http://www.komp.org) should be searched prior to undertaking any new knock-out experiment.

2. The Transgenic Mouse/ES Cell Shared Resource (TMESCSR) has the most experience using TL-1 mouse ES cells for gene targeting since they reliably give germline transmission. Since these cells were derived from 129S6 animals (previously termed 129SvEvTac) efficient gene targeting requires DNA that is of 129S6 origin for both the DNA homology arms of the targeting vector. DNA from the very closely related 129S7 (129SvEv) mice is also permitted since it is a nearly identical strain. Even so, it remains possible that small genetic differences could interfere with specific gene targeting experiments since single base mutations can prevent homologous recombination. Thus, we require that you use either 129S6 or 129S7 DNA to assemble your gene targeting vector.

   Use of PCR to amplify the homology arms is discouraged. Although this may seem like an easy way to avoid having to obtain a BAC clone even single nucleotide mismatches can greatly impair gene targeting efficiency. It may be difficult to determine if base pair mismatches identified during DNA sequence confirmation are due to PCR amplification errors or are simply 129S6 strain specific.

   a. The RPCI-22 BAC genomic DNA library was made from 129S6 ES cells (http://bacpac.chori.org/mouse22.htm ); however, this particular library was not end-sequenced so it is not indexed in GenBank or Ensembl; rather, it is available as a spotted array and clones must be identified by hybridization and then characterized to determine whether they contain the full region of interest. Nonetheless, this is the only source we know of for obtaining BAC DNA that is perfectly isogenic to the TL-1 mES cells. It is generally only necessary to screen one segment of the library to obtain several BACs containing your gene of interest.

   b. A BAC genomic DNA library has also been made from AB2.2 ES cells, which were made from 129S7 mice. The library has been end-sequenced and is indexed in Ensembl (http://www.ensembl.org). Thus, the easiest way to obtain a genomic DNA clone for gene targeting experiments is to identify a BAC clone covering a ~30-40 kb region of your gene of interest using the Ensembl browser. You can search the Ensembl web site for your gene at the Ensembl homepage. To visualize these BAC clones, click on the [ContigView] of your search result, check the box labeled ‘129S7/AB2.2 clones’ under ‘DAS Sources’ drop-down box under the Detailed View section. The 129S7/AB2.2 BAC clones are displayed as green and pink, which indicates the orientation of the DNA insert in the vector. End reads are shown as grey bars. The clones may be purchased from Geneservice, Ltd. online at http://www.geneservice.co.uk. They are described on the company’s website as Mouse bMQ BACs from the Sanger Institute.

   The TMESCSR has also developed mES cells from C57Bl6 mice. However, the germline transmission efficiency of these cells is not yet known so use of these cells is not recommended at this time.

3. It is highly recommended you assemble your gene targeting vector using BAC or Red/ET recombineering methods. This is advantageous since longer homology arms can easily be incorporated and since the method does not rely on the fortuitous presence of restriction enzyme sites. However, depending on the type of targeting vector being made, not all starting materials are available from online resources.

   BAC or Red/ET recombineering can be performed two different ways. The Stewart method has the advantage of not requiring transfer of the BAC DNA into E. coli EL350 cells, as does the Copeland method. Otherwise, the methods are similar and success is readily achievable with either approach.

   a. Stewart method. A kit containing the necessary material and information can be purchased from GeneBridges at http://www.genebridges.com/.

   b. Copeland method. Protocols and reagents are available at http://recombineering.ncifcrf.gov/.

4. All replacement-type targeting vectors require two DNA homology arms. It is highly recommended that the long arm be at least 5 kb in length, and the short arm be at least 2 kb. This is easy to achieve if BAC recombineering is used to assemble the targeting vector. If possible the homology arms should be free of repetitive DNA sequences. The absolute minimum size of the DNA homology arms is 4 and 1 kb for the long and short arms, respectively. Gene targeting will not be attempted if the arm lengths do not meet these minimum requirements. LoxP sites or introduced point mutations that interrupt a DNA homology arm must not be considered when determining the lengths of the DNA sequence. Due to the exquisitely sensitive nature of homologous recombination any such changes, even a single base, reduce the effective length of the DNA homology regions.

5. When creating a conditional allele the loxP flanked excisable region should be kept as short as possible to minimize the probability of homologous recombination occurring in this region and to be sure that Cre-mediated recombination occurs efficiently in the mouse. While there is no firm rule about the length of the LoxP-flanked region it is generally not necessary for it to be longer than 1 – 2 kb. Also, please keep in mind that Cre-mediated recombination in mice is more efficient when the LoxP sites are not too widely separated.

6. The targeting vector must include at least one chemical selection cassette. Cassettes containing both a phosphoglycerol kinase (PGK) promoter and a neomycin (G418) resistance gene are widely used for so-called positive selection. Recently, use of puromycin resistance has gained favor among some investigators. The TMESCSR can perform positive selection for neomycin, puromycin or hygromycin resistances. Many promoters that function well in somatic cells (CMV, for example) do not function in ES cells and constructs using this promoter to drive a selectable cassette will not be electroporated.

7. It is nearly unthinkable to design a targeting vector without being able to remove the positive selection cassette by site-specific recombination. This is due to numerous reports indicating aberrant effects on the expression of neighboring genes. Thus, the positive selection cassette should be flanked by recombinase recognition sites such as FRT or LoxP. If the goal is to generate a Cre-mediated conditional allele then FRT sites should flank the selection cassette. The use of a triple loxP strategy is not recommended except in special situations. A dual LoxP/FRT strategy is preferable given the high efficiency of Flp-mediated recombination by FlpE transgenic mice. It is suggested that the positive selection cassette is removed in mice versus in the ES cells. First, the intermediate allele may function as a hypomorphic allele and thus is of possible value. Second, identifying FlpE-mediated recombination in ES cells is not nearly as straightforward as in mice. Third, minimizing in vitro manipulations of the targeted cells will improve the likelihood of the ES cells transmitting through the germ line.

8. The incorporation of a negative selection cassette outside the short arm of homology is also highly recommended since this may improve the ratio of targeted clones versus random integrants, thereby decreasing the number of clones that need to be picked and screened. The use of a diphtheria toxin (DT) cassette for negative selection is preferable over the use of Herpes Simplex Virus thymidine kinase (HSV-TK) since it avoids the use of gancyclovir. However, since PGK-HSV-TK cassettes are in wide use the TMESCSR is happy to accommodate this approach.

9. DNA sequence validation of the targeting vector should be performed. At a minimum, DNA sequencing across all cloning or recombination junctions must be performed. To assure the lack of mutations any important open reading frames should also be sequenced, especially if PCR was used at any point in the assembly process.

10. All recombinase recognition sites (e.g. FRT and loxP) must be sequence confirmed, regardless of origin.\

11. The orientation of all selection cassettes needs to be known and documented.

12. A complete, assembled DNA sequence file needs to be provided to the TMESCSR prior to electroporation. The DNA sequence file should be in GenBank format and include all annotations present in the gene target. Vector NTI or other DNA applications may be used to generate this type of file. These DNA sequences will be entered into a database and will be retained as confidential information until such time that the results of the study are published by the Principal Investigator.

13. It is recommended that the screening and validation of all homologous recombination events be done by Southern blot hybridization. In our experience it is more reliable than PCR since it also detects potentially mosaic clones. The initial screen is best performed across the short homology arm. However, before any clone is used to produce mice it is essential to confirm homologous recombination at both the 5’ and 3’ ends of the targeting vector since it is not uncommon for there to be occasional other changes in the gene locus.

14. If the gene targeting vector will be used to generate a conditional allele then it is also necessary to develop PCR primers that can be used to detect the presence of the LoxP site that is located most distal from the antibiotic selection cassette. The presence of this LoxP site needs to be confirmed prior to blastocyst injection. The PCR primers developed for this purpose can also be used for genotyping of mice.

15. Careful thought should be given to the screening strategy before embarking on vector assembly. Both the cost and digestion efficiency of restriction enzymes vary greatly. It is also important to know the chromosome on which the gene you want to target resides. Traditional gene targeting efforts have not been successful on the Y Chromosome but some more advanced gene targeting methods have been successful (Simpson et al., 2002, Genesis 33: 62-66 and Rohozinski et al., 2002, Genesis 32: 1-7.) Since germline transmission after blastocyst injection involves passage of the mutation through male germ cells, genes located on the X Chromosome present unique challenges. Moreover, mutations on either sex chromosome may present other issues such as sex-linked lethality. Thus, it may be necessary to use a conditional approach to target genes on the X Chromosome.

16. All DNA probes and restriction digests must be tested by Southern blot hybridization prior to the TMESCSR performing an experiment. It is essential to use isolated from the TL-1 ES cells for these blots. The TMESCSR is happy to provide DNA isolated from these cells specifically for this purpose. Gene targeting experiments will not be initiated until the PI demonstrates the ability to detect the wild type locus by Southern blot hybridization.

17. The DNA which is used for electroporation of ES cells must be of high quality, and provided to the TMESCSR as linear DNA. Thus, it is important not to forget to include a unique restriction site somewhere outside the homology arms for linearization. Qiagen kits have a proven track record of working very well however other several other brands of kits also provide good results. Also, it should be kept in mind that linear DNA is unstable so once your DNA has been quantified and verified electrophoretically that it is linear, the sample should be stored frozen at -20 degrees C and not thawed, even in transit to the TMESCSR.

18. The design of all new targeting vectors will be reviewed by the TMESCSR Management Team prior to initiation of the experiment. The TMESCSR has no obligation to perform a poorly designed experiment. Indeed, this review is performed to avoid wasted effort and expense, and to increase the likelihood of the desired outcome. For this reason, it is highly recommended that the design of a gene targeting vector be reviewed by one of co-directors before effort is expended in assembling the vector.

These guidelines were written by Mark Magnuson with input from Trish Labosky, Doug Mortlock, Jennifer Skelton, Weiping Yuan and Jill Lindner.

Last update: June 15, 2008