Cell-Type-Specific Transgenesis in the Mouse
Since the early 1980s, when the first transgenic mice were generated, thousands of genetically modified mouse lines have been created. Early on, Jaenisch established proof of principle, showing that viral integration into the mouse genome and germline transmission of those exogenous sequences were possible (Proc Natl Acad Sci USA 71:1250–1254, 1974). Gordon et al. (Proc Natl Acad Sci USA 77:7380–7384, 1980) and Brinster et al. (Cell 27:223–231, 1981) subsequently used cloned genes to create “transgenic constructs” in which the exogenous DNA was randomly inserted into different sites in the mouse genome, stably maintained, and transmitted through the germline to the progeny. The utility of the process quickly became apparent when a transgene carrying the metallothionein-1 (Mt-1) promoter linked to thymidine kinase was able to drive expression in the mouse liver when promoter activity was induced by administration of metals. In an attempt to find stronger and more reliable promoters, viral promoter elements from SV40 or cytomegalovirus were incorporated. However, while these promoters were able to drive high levels of expression, for many applications they proved to be too blunt an instrument as they drove ubiquitous expression in many, if not all cell types, making it very hard to discern organ-specific or cell-type-specific effects due to transgene expression. Thus the need to find cell-type-specific promoters that could reproducibly drive high levels of transgene expression in a particular cell type, e.g., cardiomyocyte, became apparent. One such example is the α myosin heavy-chain (MHC) promoter, which has been used extensively to drive transgene expression in a cardiomyocyte-specific manner in the mouse. This chapter, while not written as a typical methods section, will describe the necessary components of the α myosin promoter. In addition, common problems associated with transgenic mouse lines will be addressed.
Key wordsMouse Cardiomyocyte Transgenesis Heart Muscle
Traditionally, classical genetics has involved the process of randomly mutating an organism’s genetic complement (DNA or RNA) either by radiation or mutagenic substances, followed by a systematic search for a phenotype and the associated mutation. With the development of transgenesis the process was reversed, with the genetic sequence of interest known and the transgene incorporating either the normal sequence or one in which a known mutation was made (1–3). Through micromanipulation, the modified DNA is injected into a fertilized mouse egg, where the modified DNA is inserted into the mouse genome by nonhomologous recombination. The number of transgene sequences (copy number) and the site of insertion into the host genome are random, although using specially engineered constructs (4, 5) this potential limitation can be overcome. Transgenic mice have allowed researchers to study inherited diseases, define metabolic and signaling pathways, delineate structure–function relationships in critical proteins, and determine the function(s) of newly discovered sequences. For transgenesis to be successful several technical factors need to be considered, such as the purity and concentration of the DNA to be injected, the skill of the injector, and the rigor with which animal husbandry is practiced. However, of paramount importance is the transgenic construct itself, which consists of four components: (1) The promoter, which drives transcription of the transgene to be expressed. (2) The cDNA insert, which contains the sequence encoding the peptide to be expressed. (3) The untranslated region and polyadenylation (polyA) signal lying 3′ to the cDNA and (4) The plasmid vector. Each of these components will be considered in detail in Subheading3.
Mouse genomic DNA or (purchased) Bacterial Artificial Chromosomes (BACs) containing the gene of interest.
Taq polymerase with proofreading ability designed for long templates or GC-rich regions.
Gene-specific primers with cloning sites incorporated into the 5′ end.
Reverse transcriptase or cDNA clone.
Competent bacterial cells.
Qiaex gel extraction kit.
6.3.1 Choosing a Promoter
The choice of the promoter that will be used to drive expression of the construct is a critical one. Ideally, a suitable promoter has already been defined and one can “clone by phone.” Promoters of widely varying specificity and strength have been used ranging from promoters that can drive expression in all tissues and at all developmental stages, to those that are restricted in their patterns of expression to a single cell type at a particular developmental stage. The purpose of the particular experiment will dictate which promoter is used but, in general, the more precisely expression is controlled, the more easily the investigator can interpret the resulting phenotype. This chapter will focus on precisely manipulating the output of the cardiomyocyte and thus, the murine α myosin heavy-chain (MHC) promoter will be used as an example to help illustrate the process of building a transgene (6).
If a promoter has not been previously characterized, as a first step a detailed search of Pubmed (http://www.ncbi.nlm.nih.gov/pubmed/) should be performed for the gene of interest. Often, a bioinformatics-based approach is useful and an analysis of potential binding sites may already exist. In addition, Bioinformatic Harvester (http://harvester.fzk.de/harvester/) is an excellent compiler of gene structure and function. With the complete sequence and extensive annotation of the mouse genome available, the DNA sequence surrounding the transcriptional start site can be (relatively) easily analyzed. By using the available sequence, a homology search, using Blast (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi), against other species will reveal conserved regions that may define the most important potential regulatory elements that are necessary for cell-type controllable expression. Other bioinformatic-based approaches may also be useful in designing a new promoter. For example, the program Trafac (http://www.cincinnatichildrens.org/research/cores/informatics/labs/aronow/tools/trafac.htm) screens for potential protein-binding motifs and compares potential elements between different promoters and between species, emphasizing the conserved nature of the cis transcriptional elements.
Once the sequence of the putative promoter is selected, it can now be isolated from the mouse genome. Previously, promoter regions were isolated by screening and purifying bacteriophage or cosmid clones that contained mouse genomic DNA. As shown in Fig. 1b, the mouse α MHC promoter was contained within a single restriction fragment that spanned the intergenic region between the α and the β myosin genes. In addition to the extreme 3′ terminus of β MHC, the fragment encompassed the α MHC promoter and the first three noncoding exons. The noncoding exons were left in the promoter fragment to provide a substrate for splicing, which is important for RNA processing, transport, and stability: all transgenic constructs should include a splicing site of this nature, either at the 5′ or 3′ end, or both (9, 10). Using PCR mutagenesis, a “C” was inserted into the third exon 5′ of the beginning of translation (ATG) to create a unique Sal1 site, which provides a convenient cloning site for inserting any cDNA to be expressed.
6.3.2 The cDNA Insert
The DNA to be expressed is subsequently inserted 3′ of the isolated promoter. The DNA usually consists of a double-stranded cDNA (a DNA sequence that contains only the exonic sequences) fragment cloned into a unique site behind the desired promoter, although some labs use genomic fragments that contain the coding region of the gene to be expressed, creating a minigene. In the case of a cDNA insert, PCR can be used to amplify the coding region from a cDNA template, removing as much of the 5′ and 3′ UTRs as possible. At the same time, unique restriction sites should be added onto the ends of the cDNA, allowing the fragment to be easily cloned downstream of the promoter. As an example, using the α MHC promoter, Sal1 restriction sites were added to the ends of the cDNA by PCR to aid in cloning. In those cases where the cDNA contains internal Sal1 sites, Xho1 restriction sites can be added since Sal 1and Xho 1 contain cohesive ends that can be ligated together. In those cases where the cDNA contains both internal Sal 1 and Xho 1 sites, the ends of the cDNA fragment, as well as the promoter, must be blunt ended with the use of Klenow before they can be ligated together.
In addition to incorporating restriction endonuclease sites onto the ends, the cDNA should contain sequences around the translational start site (AUG) that enhance translation. Kozak (11) performed a survey of several hundred translational start sites and discovered a consensus sequence of GCCA/GCCAUGG (the initiating methionine is in bold and underlined). To optimize translation of the transcribed cDNA, this consensus sequence should be incorporated into the construct whenever possible. Another necessary component of the cDNA is the inclusion of a termination signal at the 3′ end of the open reading frame. Any of the three stop codons can be used.
In our experience, the cDNA should contain as little of the gene’s 5′ and 3′ UTRs as possible, although minimal UTRs should be present (Subheading3.3). UTRs are known to contain regulatory sequences that may affect transcription and translation (12). miRNAs (microRNAs) are also known to interact with sequences in the 3′ UTRs and can influence both mRNA stability and translational efficiency (13). In one example of the importance of the flanking sequences, an entire cDNA with complete 5′ and 3′ UTRs was cloned into the α promoter, with no detectable protein being produced. Once the UTRs were removed from the transgene, the modified sequence resulted in detectable levels of protein.
In most cases, the cDNA PCR template can be obtained commercially or “cloned by phone.” Instead of isolating RNA, buying a kit to reverse transcribe the RNA, cloning and sequencing the potential cDNA clone, and then sorting through all of the clones that represent splicing variants, a cDNA clone can be in your lab in a few days by simply placing an order, either through commercial sources or by contacting an academic laboratory.
6.3.3 The UTR and PolyA Signal
Several studies have shown that message stability is influenced by the 3′ UTR attached to the construct (14). Transgene efficiency can be increased dramatically by using a 3′ UTR that prolongs mRNA half-life, which can increase the amount of protein produced. In the case of the α MHC promoter, the human growth hormone 3′ UTR was used. A 600-bp fragment that contains about 100 bp of 3′ UTR as well as a polyA signal was placed 3′ of the cDNA cloning site. Another useful 3′ UTR/polyA fragment is derived from the SV40 virus. The SV40 polyA fragment also contains an intron, which may be needed to confer a splicing event during processing of the primary transcript (Subheading3.1), if your promoter lacks an intron. Most commercially available plasmids use the SV40 polyA fragment, providing an easy source for ligating the appropriate signal 3′ to your cDNA as the transgenic construct is built.
6.3.4 The Plasmid Vector
All the aforementioned components are then inserted into a plasmid backbone. The vector should be a high copy number plasmid and contain an antibiotic resistance marker for easy selection. In our hands, the kanamycin resistance gene is preferable, as ampicillin can be unstable during bacterial growth. A high copy number is preferable as this lowers the total culture volume needed to produce the approximately 100 μg of construct usually requested by the Transgenic Injection Facility. However, if the construct contains repetitive sequences that have the potential to recombine during growth, a high copy number plasmid should be avoided and a low copy number plasmid backbone used instead, such as pBR322 and its derivatives.
6.3.5 Purification of the Transgene for Injection
The transgene, consisting of the promoter, cDNA, and polyA, needs to be digested away from the plasmid backbone. After digestion, the transgene needs to be purified. We normally digest 100 μg of DNA in a volume of 300 μl. The DNA fragments are separated on an agarose gel and the band containing the transgene isolated. The DNA is purified using standard column or extraction procedures: we have found that the Qiaex gel extraction method (Qiagen) provides DNA suitable for pronuclear injections. After extraction, the DNA should be resuspended in Tris/EDTA (the exact concentration and pH requested by your transgenic injection facility may vary). The DNA must be filtered or passed through a spin column. This procedure will remove any contaminating matrix or beads, which might clog the microinjection pipette. In all cases, it is a good idea to check with your transgenic facility as they may have a preferred method of transgene purification (seeNote1). Many facilities offer a DNA purification service for an additional fee and we strongly recommend that the investigator use their services as it can avoid considerable delays or finger pointing later on if founders are not forthcoming. Some facilities may add an additional step for quality control and recommend a test injection with the DNA to see if the postinjected oocyte can undergo a few rounds of division, indicating that the DNA itself is not lethal to the cell.
If there is no backlog of requests at the Transgenic Facility, then tissue samples from potential founders (transgene-positive mice) can be expected in approximately 4–6 weeks. The facility will start DNA purification within days, followed by quantification and dilution of the DNA. Within 1–2 weeks, the DNA should be injected into the fertilized mouse eggs. In about 19 days the pups will be born and after 10 days they will be large enough for sample collection (usually an earclip). The sample is digested and PCR analysis used to check for the presence of the transgene.
Use a successful Transgenic Facility. Not all facilities have a successful track record or a high success for making transgenic mice. Even the best facilities have their bad days (or weeks) when success is limited. Some though may have more bad days than good and it is important that the investigator have a clear sense of the overall competence of the facility. Many academic institutions with very competent personnel offer offsite pricing and even though it may appear to be more expensive, the amount of time saved more than makes up for the price differential that may be incurred.
No pups. If no transgenic pups can be detected from approximately 200 oocyte injections, this may indicate that the transgene encodes a protein that is lethal during embryonic development. If this is the case then more eggs need to be injected and the embryos analyzed during their development to determine the nature of the lethal event. To avoid lethality, either a different promoter that is tightly controlled such that expression occurs only after development, or an inducible system must be considered.
Mosaic founders. The transgenic facility will inject the DNA into the pronuclei of single-cell fertilized eggs. The DNA normally integrates at this stage and the transgene will be present in all cells of the mouse, including the germline. However, if integration occurs at a later stage when there are two or more cells, the resultant mouse may be mosaic for the transgenic DNA, with the transgene present (and expressed) in some of the cells, but not in others. It is possible that no, or very few, F1-generation transgenic mice will be born from the mosaic founder if those progenitor cells that form the germline did not acquire the DNA. If no transgenic offspring from a positive founder are obtained in the initial litter, we routinely screen at least three additional litters from the potential founders for positive offspring before terminating the line.
Transgene expression levels and molecular torture. Most transgenes have some degree of sensitivity to the surrounding chromosomal context and/or some degree of copy number-dependent expression. As such, not all lines containing the identical transgenic construct will express the transgene at the same level, as each line may contain varying (1→100) copies of the transgene inserted at different sites in the mouse genome. This can be useful as a range of protein expressions can illuminate the physiology of the resultant phenotype. However, nonphysiologically high levels of expression can be misleading, resulting in cell responses that are related simply to the abnormally high protein level rather than being directly related to the protein’s function. If a mutated protein is being expressed this can be controlled by making transgenic lines in which the transgene consists of the normal protein being expressed at levels at or above the mutant polypeptide. In cases where this control cannot be done, the data from the transgenic line must be cautiously interpreted.
Animal husbandry. Most if not all injection facilities will use mice that have been raised in barrier rooms and housed in microisolator cages: these mice are referred to as being “clean.” That is, they have been purchased from a reputable provider, or bred in house and have been tested serologically for viruses and pathogens. NIH mandates that transgenic mice made under the auspices of government support be freely shared. By keeping some of the resultant transgenic mice in a barrier, or clean room, mice can be easily exchanged between facilities without spreading disease. Additionally, some animal models will require that the mice remain clean, since viral or bacterial contamination will complicate the phenotype. As examples, results will be compromised when studying liver function if the mice have hepatitis, or when studying digestive diseases and the mice have helicobacter.
Sex-linked transmission . The genomic integration of the transgene is a random event and occasionally the DNA inserts into the X or Y chromosomes. An insert in the Y chromosome will be obvious when no transgene-positive females are detected in the first- or second-generation offspring. An insertion into the X chromosome may be difficult to interpret if the transgene is silenced by x-linked inactivation in which some cells are expressing an active transgene, while other cells are not. These lines should be discarded.
Double insertions. Chromosomal integration most frequently occurs at a single site. In a few instances the transgene will insert into two distinct sites, possibly on different chromosomes. Sometimes the offspring will contain both inserts; at other times the pups may contain one or the other insert. This will become apparent when a single litter gives a wide range of expressions, or if too many positive pups are obtained (nonmendelian inheritance). Since maintaining the double insertion on one transgenic line is difficult, the two insertions should be bred out into two separate sublines.
Insertional mutagenesis. During the recombination events that underlie transgene insertion, genomic deletions may occur at the integration site. The deleted DNA may range from a few bases to a kilobase in size. Normally, as long as the transgenic mice are maintained as heterozygotes, the deletion will not be noticed as the mutation is only rarely dominant and one intact allele is still present. However, if the transgene is bred to homozygosity then the (potential) effects of the deletion may become apparent (15), with a resultant, confounding phenotype. For this reason it is always a good idea to compare data using at least two distinct lines (in case the mutation is, in fact, dominant) and only use outbred transgenic offspring to ensure that the mouse is heterozygous for the transgene.
- 9.Illicit splicing events. In a few instances one is able to detect mRNA from the transgene, but no transgenic protein is produced. Using the α MHC promoter, after analysis by RT-PCR we determined that the RNA had been improperly spliced. Instead of the expected splicing between the second and third noncoding exons of the promoter (Fig. 3), the second exon was spliced into the cDNA, removing a third of the coding region. The third exon was later removed and the cDNA ligated directly to the second exon in an attempt to eliminate the errant splicing. The new transgene spliced from the first exon into the cDNA, again removing most of the coding sequence. In another attempt we changed the sequence of the splice site in the cDNA. The construct still spliced incorrectly. Analysis of the cDNA sequence revealed several potential splice sites, which were eliminated where possible by changing the cDNA sequence. Most amino acids are encoded by more than one DNA triplet, allowing the sequence to be modified while conserving the proteins’ amino acid sequence.
Lack of a phenotype . In some transgenics no phenotype is observed. Lack of a phenotype most often means that the investigator has not devised a suitable screening procedure or has a predetermined notion of what the phenotype should be. Occasionally, however, this can be due to a complete lack of protein function. This can arise for a number of reasons. The protein may not be trafficked correctly within the cell. If a “sorting” signal is left off of the transgene, such as a nuclear localization signal, then the protein may not reach the correct cellular subcompartment. Another possibility is that the transgene is post-transcriptionally controlled or the protein is post-translationally modified and subsequently degraded. The most frequent cause is due to a mutation occurring in the transgene, giving a truncated protein product or no protein at all. Your construct should be sequenced at all steps of development and care should be taken to derive the transgene from the mouse strain that will be the source of the oocytes used for injection, to minimize strain differences in the DNA that might result in subtle changes in protein sequence (seeNote 13).
Strain differences. The Mouse Sequencing Consortium recently finished sequencing the genome of the C57BL/6J strain of mouse (http://www.genome.gov/10001859) with genetic analysis on fifteen of the most commonly used strains currently underway. Strain variations are significant and must be considered as they can have a major impact on phenotype presentation. As an example, several strains of mice were characterized as to their cardiac function, with statistically significant differences (16). In a transgenic model, strain differences were observed between C57 and FVB mice when a GATA4 mutation was introduced (17), leading to variations in valve formation. Most journals will ask you to list which strains were used and the percentage of each if the strain is not pure. Some journals actually go so far as to demand that experiments be conducted in a pure, defined strain.
Phenotypic consequences of reporter gene expression or epitope tagging. Reporter genes are sometimes used in transgenesis as a way of easily assessing the cellular/tissue location of where a promoter is active or if a certain event occurs, such as genetic recombination. One common marker is GFP (Green Fluorescent Protein), which is often linked to a transgene, creating a fusion product. The GFP serves as a useful marker, allowing your protein to be tracked as it moves about the cell, or to quantify the levels of expression. An area of concern is that GFP can affect biological function, e.g., possibly altering the electrophysiology of cardiac cells (18). Another protein that may cause a phenotype is a component of the commonly used binary system, the Cre protein in the Cre-lox system. Cre expression can, for example, cause dilated cardiomyopathy when expressed at high levels (19). Another binary system, the tetracycline inducible promoter, uses a transactivator protein, which when expressed at high levels causes a cardiomyopathy (20, 21).Frequently, epitope tags are added to the coding portion of the transgene, allowing the protein of interest to be tracked or identified within the cell by immunohistochemical means. These extra bases are usually incorporated onto either the N- or C-terminus of the protein. In some instances the tags can interfere with protein function by blocking protein-protein interactions or affecting intramolecular folding. Most tags contain charged amino acids or are highly structured, making them great epitopes but (potentially) terrible additions to your protein in terms of its normal function. A review by Maue (22) lists the most commonly used tags and their amino sequences, and describes the successes and failures of adding tags to various constructs.
Phenotype drift. Over time, an observed phenotype may change or drift within a colony of transgenic animals. For example, an investigator might report that all transgenic mice died at 6 months of age but, 2–4 years later, note that all are now living consistently for 9–12 months. One explanation is that over time those mice that are the healthiest have a tendency to breed more frequently and for a longer period of time, skewing the colony’s phenotype, while the sicker mice become progressively under-represented. There is no easy way to avoid this since the breeding ability of the mice is not known in advance, but the cautious investigator will only use breeding pairs that are less than 6–7 months and continuously monitor the phenotype of the colony.
Transgene silencing. Genes can be silenced by epigenetic changes, such as DNA methylation and histone acetylation. DNA becomes methylated when enzymes add methyl groups to the cytosine residue of CpG sequences (23, 24, 25). This pattern of methylation is then inherited by the next generation of mice and can lead to transgene inactivation. One can minimize the problem by using endogenous sequences for all elements of transgene construction, but it remains a potential issue despite such precautions.
Transgene instability. Some genomic sequences are known to be unstable, especially those that contain repetitive elements. If repetitive elements are included in your construct or generated during the insertional event, rearrangements may occur, in some instances deleting the transgene and some of the surrounding chromosome.
- 14.Knouff, C., Malloy, S., Wilder, J., Altenburg, M.K. and Maeda, N. (2001) Doubling expression of the low density lipoprotein receptor by truncation of the 3′-untranslated region sequence ameliorates type iii hyperlipoproteinemia in mice expressing the human apoe2 isoform. J Biol Chem, 276, 3856–3862.PubMedCrossRefGoogle Scholar
- 18.Sekar, R.B., Kizana, E., Smith, R.R., Barth, A.S., Zhang, Y., Marban, E. and Tung, L. (2007) Lentiviral vector-mediated expression of GFP or Kir2.1 alters the electrophysiology of neonatal rat ventricular myocytes without inducing cytotoxicity. Am J Physiol Heart Circ Physiol, 293, H2757–H2770.PubMedCrossRefGoogle Scholar
- 20.McCloskey, D.T., Turnbull, L., Swigart, P.M., Zambon, A.C., Turcato, S., Joho, S., Grossman, W., Conklin, B.R., Simpson, P.C. and Baker, A.J. (2005) Cardiac transgenesis with the tetracycline transactivator changes myocardial function and gene expression. Physiol Genomics, 22, 118–126.PubMedCrossRefGoogle Scholar