General or Cell Type-Specific Deletion and ... - Semantic Scholar

METHODS 20, 205–218 (2000) doi:10.1006/meth.1999.0938, available online at http://www.idealibrary.com on

General or Cell Type-Specific Deletion and Replacement of Connexin-Coding DNA in the Mouse Martin Theis, Thomas M. Magin, Achim Plum, and Klaus Willecke 1 Institut fuer Genetik, Abteilung Molekulargenetik, Universitaet Bonn, Roemerstrasse 164, D-53117 Bonn, Germany

Here we describe several gene targeting approaches currently used in our laboratory for the generation of deletion or replacement mutants of connexin genes in the mouse and discuss the advantage of the double-replacement strategy for the generation of conditional mutants. For the analysis of complementary functions of connexins, it will be necessary to generate mice with mutations in several connexin genes. We also report how this can be effectively accomplished. The replacement of targeted connexin-coding DNA with a reporter gene, to mimic expression of the deleted gene product, is currently being used in several laboratories. The use of different reporter genes or their differently localized gene products could allow distinction of promoter activity in double or triple connexin mutant mice. © 2000 Academic Press

The gene family of mouse connexins consists of at least 15 members that code for protein subunits of gap junctional intercellular channels (1– 4). Gap junctions are conduits that mediate intercellular diffusion of small molecules such as ions, amino acids, small peptides, metabolites, and second messengers (1, 5). The functions of connexin channels are supposed to include sychronization of cells in a given tissue or organ (6), growth control (7), metabolic coupling (8), and electrical coupling of excitable cells (9, 10). These hypothetical functions of connexin channels can be experimentally verified by analysis of targeted mouse mutants obtained after homologous recombination in embryonic stem (ES) cells (11, 12). Targeted gene manipulation can be accomplished by the use of replacement or insertion vectors (13, 14). Here, we focus on gene targeting using replacement 1 To whom correspondence should be addressed. Fax: ⫹49228 734263. E-mail: [email protected].

1046-2023/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

vectors (schematically outlined in Fig. 1). Connexins are particularly suited for gene targeting, because they have a relatively uniform and simple gene structure. Most of the connexin genes include one relatively short exon separated by a large intron from the second exon that encodes the entire coding region and polyadenylation signals. The coding region of most connexin genes is relatively compact (about 1 kb in length) and, thus, can easily be replaced by other DNA sequences. Until now eight connexin genes have been inactivated in the mouse, leading to a broad range of phenotypes (11, 12, 15), including embryonic lethality. To avoid embryonic lethality and to study the function of the corresponding genes in adult mice, conditional gene ablation using the Cre/loxP system (16, 17) has been worked out as the method of choice. The pioneering work of Sauer, who established the bacteriophage P1-derived Cre/loxP system in eukaryotic cells (18, 19), and of Gu et al. (20), who first used this system for homologous recombination to generate conditional mouse mutants, led to the widespread use of the Cre/loxP system in gene targeting experiments. It allows conditional gene modifications to elucidate their function in specific cell types or at defined points in development. The first component of the system is a DNA sequence (gene or part of it) flanked by two recognition sequences (loxP sites) for the recombinase Cre. Often the loxP sites are introduced by homologous recombination to generate an allele with the loxP flanked (“floxed”) DNA region and an additional “floxed” selectable marker gene, which can subsequently be removed by transient transfection of targeted ES cells (Fig. 1d) (20, 21). Mice that carry floxed alleles are assumed to be phenotypically indistinguishable from those carrying wild-type alleles. The second component, the cre gene, confers tissue specificity of the gene deletion if expressed under control of a suit205

206

THEIS ET AL.

able cell type-specific promoter (22). This tissue specificity can be combined with temporal control by a promoter with a defined developmental onset of expression (23). Wider versatility can be obtained by the use of inducible promoters (24, 25) or posttranslationally inducible Cre/hormone receptor fusion proteins (26 –29).

In several cases, the effects of connexin ablation became obvious in only a subset of organs or cell types in which this gene was expressed (30 –33). This points to functional compensation among connexins. On the other hand, despite their overall structural similarity, the channel properties of connexins are different from each other, suggesting physiologically relevant special-

FIG. 1. Gene targeting strategies using replacement vectors. (a) Single-step gene targeting. (b) Double-replacement strategy. (c) Use of Cre/loxP technique for gene targeting. (d) Use of Cre/loxP technique for generation of floxed alleles. Big black box: coding sequence of connexin; small black boxes: 5⬘- and 3⬘-untranslated regions; hatched box: marker DNA or mutated connexin-coding sequence; black arrowhead: primer for PCR analysis of transient transfection with Cre-encoding plasmid; PCR product: amplified DNA sequence indicative of homologous recombination event; small black arrow: external primer for detection of homologous recombination event; small light arrow: internal primer for detection of homologous recombination event; hatched arrowhead: loxP site. * This clone can be used for generation of a connexin-deficient mouse. Primer pair ‘b’ might lead to a large amplicon containing the selection marker DNA.

TARGETED DELETION AND REPLACEMENT OF CONNEXIN DNA

ization (34, 35). To investigate the extent of specialization or the extent of functional complementation among connexins, we have used the doublereplacement strategy (36) to exchange one connexin coding region for another one (Plum et al., unpublished). For other genes, the use of two sequential targeting steps has allowed the introduction of subtle mutations without residual selection marker at the modified locus (see Fig. 1b) (37–39). Here, we focus on the use of hypoxanthine phosphoribosyltransferase (HPRT)-deficient ES cells and HPRT minigenes for selectable homologous recombination (36, 40). In this experimental system, an HPRT minigene is used first to rescue the deficiency concomitant with the deletion of the gene of interest. Only ES cells that harbor an HPRT minigene can grow in HAT (hypoxanthine, aminopterin, and thymidine containing) selective medium. After the first round of correct gene targeting, the HPRT minigene is replaced by a modified version of the original gene or by a related sequence. In the second step, selection with 6-thioguanine allows cell survival only on loss of the HPRT minigene. Several similar approaches, using marker genes that can be selected in a positive and negative manner (36, 41) or a combination of two selectable marker genes (38, 39, 42– 44), have been described. However, this replacement of one connexin by another might be lethal or detrimental to reproduction. With the Cre/loxP system, conditional replacement of genes has become feasible (45– 47). The use of a silent, but activatable lacZ gene expressed on deletion of a floxed gene has been proposed by van der Neut (45). Such a mouse might serve as a first model system for the generation of conditional gene replacements to assess the leakiness of the constructed allele by monitoring ␤-galactosidase activity in the nonreplaced state. In this case, a recombinant allele needs to be generated that consists of a floxed connexin coding region followed by the coding region of another connexin gene (Fig. 2). Cre-mediated deletion of the floxed DNA segment leads to replacement of the floxed coding region by the second coding region. Concomitant with the floxed connexin region, a stop signal that prevents translation of the second coding region is deleted. This stop signal could be a polyadenylation signal as in several reporter transgenes (48) or the stop codon of the floxed connexin coding DNA (M. Theis, unpublished results). Tissue specificity or temporal control of replacement can be achieved similarly as with conditional ablation. To fully elucidate the extent of mutual influence or compensation among connexins, it will be necessary to generate double or triple connexin-deficient mouse mutants. Some of the connexin gene deletions are lethal and hence have to be conditional, to generate surviving mice. For this purpose, in addition to the two or three

207

connexin genes involved, at least one additional locus containing the cre transgene needs to be crossed in, which further complicates the breeding scheme. To decrease the complexity of such matings, we propose the targeted replacement of one connexin coding region by that of a posttranslationally inducible Cre recombinase (27, 28) (Fig. 2). In this case, the same connexin coding sequence that has been replaced by a cre/ hormone receptor fusion gene should be flanked by loxP sites. On induction, the deletion of this particular connexin will occur in every cell where it is expressed at the time of induction. The first connexin deletions by a reporter gene reflecting transcriptional activity of the connexin promoter have been generated (15, 32). In connexin double-deficient mice, two different reporter genes could be used to distinguish reporter gene activity. Several connexins are expressed early in development (49, 50), some of them in ES cells (30, 51, 52; A. Plum, unpublished). In these cases, different reporter genes could be tested, after appropriate replacement of connexin coding regions, already on the ES cell level. We have used the ES cell line HM-1, which is HPRT deficient and can be cultured without feeder cells (40). This feature simplifies ES cell culture, since there is no need to prepare feeder cells and culture them in parallel. In addition, there is no need to separate ES cells from feeder cells for DNA isolation, karyotyping,

FIG. 2. Possible connexin gene modifications that can be generated by the double-replacement gene targeting strategy. Big black box: connexin coding DNA; small black boxes: 5⬘ and 3⬘-untranslated regions; hatched dark gray box: altered or related connexin-coding sequence; hatched light gray box: nonconnexin sequence; hatched arrowhead: loxP site; lbd: coding sequence of ligand binding domain of hormone receptor.

208

THEIS ET AL.

Northern and immunoblot analysis, immunofluorescence, in situ X-Gal staining, or transient transfection experiments. The HPRT deficiency makes it possible to use the double-replacement strategy for gene targeting. Here we describe our experience and further possible applications of the HM-1 cell line for targeted manipulations of connexin genes.

DESCRIPTION OF METHODS Culture of HM-1 ES Cells HM-1 ES cells are derived from the mouse strain 129P2/OlaHsd (53). They are grown on gelatinized tissue culture plastic in an atmosphere containing 5% (v/v) CO 2 and do not require feeder cells. For routine culture, cells should be passaged twice a week by diluting them between 1:2 and 1:6, depending on the density of the culture. Medium (see Materials) should be changed every other day as long as the density is low, but should be changed daily when cultures become dense. The cells should not be plated too sparsely, as this favors excessive differentiation. On the other hand, like other ES cells, they should not become overconfluent. Any drastic change in cell doubling times or cell morphology is indicative of unwanted damage to the cells, e.g., chromosomal aberrations. Typical doubling times are about 22–24 h. To retain their frequency of germline transmission, it is essential to keep the passage number of HM-1 cells as low as possible. High frequencies of germline transmission strongly depend on (1) cell culture materials (i.e., tested sera, media, plasticware), (2) low passage numbers, and (3) good husbandry. We use plastic material for cell culture, cell culture-grade water, and reagents of highest purity. We test all batches of sera at concentrations up to 30% for cytotoxicity, cell doubling time, and normal morphological appearance of HM-1 cells. Sera that yield low germline transmission are discarded. Thawing and Plating of Cells Gelatinize cell culture flask with 3–5 ml of gelatin per 25-cm 2 flask and 7–10 ml per 75-cm 2 flask. Leave at room temperature for at least 10 min. Discard supernatant before plating the cells. Prepare centrifuge tube with 10 ml of prewarmed medium. Thaw cells quickly in 37°C bath and transfer into prewarmed medium. Spin for 2 min at 1200 rpm. Remove medium, replace with 5 ml of fresh medium containing penicillin/ streptomycin, and plate all cells on 25-cm 2 flask. On the next day, replace with fresh medium without antibiotics. Routine culture without antibiotics is strongly recommended. For subculturing, first gelatinize new flask(s). Aspirate the medium and wash cells, first with

5 ml prewarmed phosphate-buffered saline (PBS) and then with 1 ml ES cell–trypsin (prewarmed) per 25-cm 2 flask. Use 10 ml PBS and 2 ml ES cell–trypsin per 75-cm 2 flask. Remove and replace with fresh trypsin and leave in incubator at 37°C until cells can be dislodged by gentle agitation. This should occur after about 2 min but may take up to 5 min. Do not trypsinize too long, as this can damage the cells irreversibly. Add 1 ml of medium per 1 ml of trypsin and pipet up and down against the wall of the flask about 10 –20 times. Large cell clumps must be avoided, as they will stimulate differentiation of the cells. Check suspension under microscope before transferring cells into a centrifuge tube. Spin for 2 min at 1200 rpm (room temperature), aspirate off supernatant, and resuspend cell pellet gently in a few milliliters of medium. Replate at desired density and add 6 – 8 ml of medium to a small flask as well as 25–30 ml to a large cell culture flask. A 25-cm 2 flask yields about 7–10 ⫻ 10 6 cells and a 75-cm 2 flask about 2.5–3 ⫻ 10 7 cells. Freezing of ES Cells Freeze cells from near-confluent flasks. One aliquot should contain 2– 4 ⫻ 10 6 cells (i.e., one can obtain two or three aliquots from a small flask and five to eight aliquots from a large flask). It is beneficial to freeze cells as concentrated as possible, e.g., in 0.5–1 ml per vial. Trypsinize cells as usual. Resuspend them gently in 1 vol of normal medium. Add 1 vol of twofold concentrated freezing medium dropwise. Freeze cells stepwise by placing the vials first at ⫺20°C for 1 h, and then at ⫺70°C overnight. Cells can be stored at ⫺70°C for up to several weeks. For long-term storage, transfer cell vials to liquid nitrogen. Alternatively, place vials immediately into the gaseous phase of a liquid nitrogen container for slow temperature decrease. The following day, store in liquid nitrogen. Electroporation of ES Cells (First Gene Targeting Step) The conditions recommended will work with any ES line and are relatively insensitive to changes in cell number and amount of DNA. They do not yield the largest possible number of transformed cells but are relatively gentle to the cells.

Preparation of DNA Currently we achieve the best results with DNA purified on Qiagen Endofree (Qiagen, Hilden, Germany) columns. After linearization of plasmid DNA with a suitable restriction enzyme, extract DNA with phenol/chloroform, precipitate with ethanol, and wash twice with 70% (v/v) ethanol (endotoxin-free Qiagen). At a clean bench remove the supernatant, air-dry, and resuspend the DNA immediately before transfection. We have obtained good results omitting the phenol/ chloroform extraction, which is more consistent with the use of endotoxin-free materials.

TARGETED DELETION AND REPLACEMENT OF CONNEXIN DNA

Electroporation For a standard gene targeting experiment, trypsinize about 2 ⫻ 10 7 cells. Resuspend ES cells in 0.8 ml of HBS buffer (see Materials). Add cells to 200 ␮g linearized DNA (a typical targeting vector of approximately 10 kb) in 100 ␮l TE buffer (endotoxin-free Qiagen) in Eppendorf tube and mix. Transfer suspension into an electroporation cuvette with an electrode spacing of 0.4 cm. Pulse at 800 V and 3 ␮F using a Bio-Rad Gene Pulser (Bio-Rad, Muenchen, Germany) and incubate for 10 min at room temperature. Add cells to appropriate amount of medium, depending on the number of dishes to be used. Plating and Selection Plate about 1.5–2 ⫻ 10 6 cells per 10-cm gelatinized Petri dish (do not use flasks) in 10 ml of medium. In this calculation, cell death due to electroporation is already taken into account. Start positive selection with HAT medium or neomycin (G418; 350 ␮g/ml active compound) on the following day (Day 1). If negative selection is applied, start it at Day 3 by changing medium. For herpes simplex virus thymidine kinase (HSV-TK) negative selection, use gancyclovir at a final concentration of 2 ␮M. Leave one plate without negative selection to determine the enrichment achieved with positive–negative selection. After 3–5 days, change medium. Colonies will appear after 8 –10 days. Isolation and Expansion of Clones Passage colonies when they are about 2–3 mm in size by dislodging them from the Petri dish, using yellow tips and pipet adjusted at 150 ␮l. Transfer colonies to a gelatinized well of a 24-well plate and pipet up and down a few times to break up cell clumps. If the clones are screened by polymerase chain reaction (PCR), remove an aliquot for analysis. Avoid growth of cell clumps in 24-well plates. If this occurs, dissociate cell clumps by trypsinizing them and replate in the same well. When the cells become confluent, transfer them to a 12-well plate. Further passage the cells onto a 6-well plate and finally onto two 25-cm 2 flasks: one for preparation of genomic DNA and karyotyping, and the other to prepare two or three aliquots of frozen cells. Alternatively, when passaging from a 6-well plate to a 25-cm 2 flask, replate 20% of the cells from the 6-well aliquot onto a new 6-well plate for DNA preparation. This minimizes the number of cells taken from the expanding clone for characterization and facilitates further processing of the DNA for analysis, since one 6-well plate carrying six individual clones is easier to handle and smaller amounts of reagents are required, compared with flasks. The DNA recovered from a confluent well is sufficient for four or five restriction digests. During the whole expansion procedure keep passage number as low as possible and make sure to

209

trypsinize cells efficiently. Maintain selection until targeted ES cells are expanded and characterized by Southern blot hybridization. To remove cells from HAT selection, replace HAT with HT medium (containing hypoxanthine and thymidine) for at least 2 days before growing cells in normal medium. Cells will grow slightly faster in the absence of selective conditions.

PCR Analysis It is strongly recommended that a PCR assay be established for searching the recombined allele. For technical details of several PCR strategies see (47, 54); for general information on PCR conditions see (55). One DNA primer should bind to a sequence that is part of the introduced DNA and should not bind to sequences of the original allele (Fig. 1a). The other primer should bind outside of the homology region of the targeting vector. Only on homologous recombination should both primer binding sites occur in cis orientation at an appropriate distance and efficient amplification take place. If the shorter homology arm of the targeting vector exceeds 2 kb, however, the establishment of reliable PCR conditions may be very labor intensive. We generally establish a PCR assay by use of a test vector. Ten femtograms of test vector DNA should be sufficient to detect the expected PCR product in a background of wild-type genomic DNA. The test vector contains a stretch of DNA identical to the sequence newly generated by homologous recombination and providing the primer binding sites. Spin down aliquots of the cells removed after breaking up the colonies, discard supernatant, and resuspend pellet in 40 ␮l PCR buffer-containing detergent and proteinase K (final concentration 200 ␮g/ml). Incubate for 1 h at 56°C and 15 min at 95°C to inactivate proteinase K. Use 10 ␮l for PCR (note that the DNA is already dissolved in PCR buffer). Second Step of Gene Targeting (Double Replacement) Using HPRT Minigenes Transfer cells in standard medium as described above. Electroporate about 2 ⫻ 10 7 cells, as before, and plate onto six 75-cm 2 flasks. To ensure that the successfully recombined cells have lost the HPRT activity, wait 6 days after electroporation before applying 6-thioguanine (6-TG, 2 ␮g/ml) selection. Most likely, it will be necessary to reduce the cell density. If necessary, dilute cells so that the flasks are almost confluent on Day 6. Then, trypsinize cells and plate 1–1.5 ⫻ 10 6 cells from each of the six different flasks onto 10-cm dishes, in medium containing 6-TG. Change medium after 2 and 6 days. About 1 to 10 colonies per dish will become visible by eye around Day 10. Pick, expand, and analyze the clones as described. Use a PCR assay suitable for the detection of homologously recombined clones and discard PCR-negative clones. Record the

210

THEIS ET AL.

origin of the picked ES cell clones, since clones from different initial flasks are due to independent recombination events, whereas clones from the same flask might be derived from the same cell. Transient Transfection with Cre-Encoding Plasmid DNA for Partial Deletion of Recombined Allele with Three loxP Sites This protocol is designed for isolation of ES cell clones that have undergone partial deletion of a selection marker flanked by loxP sites (Fig. 1d). The use of plasmids expressing cre under the control of relatively weak promotors such as pMC-Cre and pIC-Cre (47) should lead to decreased activity of Cre in at least some of the transfected cells. Too high activity might lead to total deletion in most of the recombinant clones.

Electroporation Prepare DNA as described for targeting constructs, but do not linearize. Check by agarose gel electrophoresis that the plasmid DNA has retained the supercoiled conformation. Use the same conditions as for electroporation with the targeting vector, but take the smaller size of the Cre-encoding plasmid into account when calculating the amount of DNA for the transfection. Plating and Selection After transfection, plate 3– 4 ⫻ 10 6 cells per 10-cm Petri dish. Replate 48 h after electroporation at a density of 0.5–5 ⫻ 10 5 cells per 10-cm dish. Delaying the final plating allows the cre plasmid to be diluted out in the dividing cell population, giving rise to partial deletion in cells with a sufficiently small amount of Cre recombinase. The delay also circumvents the problem of mixed colonies, where some of the cells and their descendants have undergone deletion, while others show no deletion at all (47). It is not necessary to replate all electroporated cells. Most of the cells can be frozen for later analysis, if selection has failed (47). If a floxed neomycin phosphotransferase/thymidine kinase (neo-tk) construct has been used, it is recommended that selection be started with 2 ␮M gancyclovir 48 h after replating and untransfected cells be used as a control to assess the efficiency of selection. If the targeting vector contains a floxed neo sequence, it is strongly suggested that duplicate plates of the picked clones be prepared to test for neomycin sensitivity after removal of the floxed neo sequence as described in (47). PCR Screening of Deleted Clones It might be preferable to design a PCR assay to test the clones for deletion of the floxed selection marker. Several strategies are possible (Fig. 1d). a. Screening for Absence of PCR Signal The use of internal primers for the selection marker will lead to a signal if the selection marker DNA is still

present in the cell clone even if it is a mixed colony. The corresponding PCR product will be formed and the clone can be discarded. However, it is not possible to distinguish clones that have undergone total deletion from clones that exhibit the desired partial deletion. b. Screening for Presence of PCR Signal Use primers spanning the floxed selection marker DNA. One primer binding site should reside within the floxed region of the connexin gene but outside of the floxed marker region; the other primer should bind to an endogenous sequence that is part of the homologous region during the initial targeting step for introduction of the floxed allele. Only on deletion of the floxed selection marker, can a DNA fragment be detected representing the desired partial deletion. When establishing this kind of PCR, one should take into account that primers spanning the floxed selection marker will also amplify the wild-type allele. The wild-type amplicon and the floxed amplicon differ in size only slightly due to the insertion of the loxP site (34 bp) and, possibly, additional base pairs of the multiple cloning site from which the loxP site had been excised during construction of the targeting vector. Often the difference is about 100 bp, which makes distinction between wildtype and floxed alleles feasible, if the amplified floxed fragment is in the size range 500 –1000 bp. If total deletion has occurred, only the wild-type fragment, not the floxed amplicon, can be detected. One disadvantage of this strategy is that a mixed colony, with some cells still unrecombined, cannot be distinguished from a homogeneous cell colony exhibiting only the desired partial deletion. An advantage is that there is always an internal control for each PCR, namely, the amplification of the wild-type allele. c. Alternative Screen for Presence of PCR Signal A third PCR could span the region of genomic DNA containing the separate loxP site that flanks, together with the selection marker sequence, the connexin gene. In the heterozygous ES cells, two amplified fragments of different length are generated: one wild-type amplicon and a floxed amplicon that is longer due to insertion of the loxP site. Again, the length difference will be at least 34 bp. Often the loxP-containing fragment, inserted into the targeting vector, is approximately 100 bp. In ES cells that have undergone total deletion on transient transfection and in wild-type revertants, only the shorter wild-type amplicon is detected. Cell clones negative for the marker DNA but positive for the floxed amplicon are good candidates for the desired partial deletion. After expansion and DNA isolation for Southern analysis, it is recommended that the possible random insertion of the cre plasmid DNA used for transient transfection be ruled out by PCR.

TARGETED DELETION AND REPLACEMENT OF CONNEXIN DNA

Functionality of loxP Sites The functionality of the two residing loxP sites should be tested with an additional transient transfection before blastocyst injection (56), as partial deletion might have been due to a mutation in the Cre recognition sites, resulting in a floxed allele with nonfunctional loxP sites. After transient transfection with cre, as described above, screen colonies by PCR as outlined before for partial deletion. This protocol is also suitable for deletion of floxed selection marker DNA after replacement of an endogenous DNA sequence by a modified sequence (Fig. 1c) (57). Only two loxP sites are introduced into the recombined allele. In this case, the use of a Cre-encoding plasmid with a stronger promoter such as PGKCre NLSbpA (47) is recommended to enhance efficient deletion of selection marker DNA. However, to reduce culture time and retain germline competence, it is preferable to remove floxed selection marker DNA in the mouse by mating to a deleter mouse (58). Deleter mice harbor a cre transgene that is expressed early in mouse development, leading to systemic deletion of floxed DNA sequences. Alternatively, EIIa-cre mice (59) or PGK-Cre m mice (60) can be used. Isolation of Genomic DNA and Southern Analysis

Preparation of Genomic DNA The DNA isolation procedure used in our laboratory is a modified version of the isolation protocol described in (61). Grow ES cells to confluence in a 25-cm 2 flask. Wash cells twice with 5 ml PBS, remove PBS, add 2 ml lysis buffer containing 200 ␮g/ml proteinase K (see Materials) that has been prewarmed for 5 min, at 65°C, agitate for 30 min on a rocking platform, transfer lysate to a 2-ml tube, and incubate for at least 24 h at 55°C in a roller incubator. Alternatively, wash confluent six-well plate twice with PBS (2 ml per well), remove PBS, add 1 ml prewarmed lysis buffer, agitate at room temperature for 30 min, seal the six-well plate with Parafilm, and incubate at 55°C for at least 24 h. For large-scale analysis of recombinant clones, add 1 vol 2-propanol and mix until a cloudy precipitate becomes visible. Do not add additional salt! Transfer the precipitate to 70% ethanol using a yellow pipet tip. Then transfer precipitate to an empty tube, air-dry, and dissolve DNA in 200 – 400 ␮l (25-cm 2 flasks) or 150 ␮l (six-well plate) TE buffer, pH 7.5 (62). For the detailed analysis of candidate clones for blastocyst injections, phenol/chloroform extraction (62) using larger amounts of cell lysates is recommended prior to isopropanol precipitation (in 75-cm 2 flasks). Southern Blot Analysis The restriction digest of genomic DNA is carried out as described in standard protocols (62). About 15 ␮g of

211

genomic DNA is digested overnight with 40 –50 U of the restriction enzyme in a volume of 40 –50 ␮l in the presence of 100 ␮g/ml RNase. Gel electrophoresis, capillary transfer, and hybridization are carried out according to standard protocols (62). Karyotyping of ES Cells Before using a clone for a second electroporation or for blastocyst injections, the number of chromosomes should be determined by karyotype analysis. A high degree of euploidy is necessary but does not guarantee germline competence. Therefore, it is preferable to characterize a sufficient number of euploid clones to increase the probability of germline transmission. Starting with confluent 25-cm 2 flasks, the cells are arrested in metaphase by incubating them for 1 h at 37°C, 5% CO 2 with 1 ml fresh medium containing 10 ␮l Demecolcine solution (10 ␮g/ml in Hanks’ balanced salt solution; Sigma D-1925; toxic— handle and dispose according to safety guidelines). Remove medium and wash twice with PBS. Add 1 ml of trypsin solution and incubate for 4 min at 37°C, stop trypsinization by adding 5 ml medium, spin down cells, remove medium, and add 1 ml of 0.56% KCl dropwise to the cell pellet. Resuspend cells by flicking, then add another milliliter of 0.56% KCl and mix by flicking. Incubate at room temperature for 8 –10 min. Spin down gently (700 rpm, 5 min) and remove supernatant. Add dropwise 2–3 ml of chilled fixation solution (methanol/acetic acid 3:1, freshly prepared) onto the pellet, resuspend by flicking, incubate 5 min at room temperature, spin down gently for 5 min at 700 rpm, and remove supernatant. Repeat fixation procedure once or twice. Finally resupend pellet in 0.5–1 ml fixation solution, drop the solution onto clean slides from a distance of 15–30 cm, and air-dry slides. Check under the microscope if the broken nuclei are sufficiently separated from each other to ensure that every chromosome can be assigned to its metaphase spread. Incubate slides for 1 min in Giemsa staining solution (Giemsa stain, modified, 0.4%, Sigma). Wash twice with distilled water, air-dry, and mount the slides. Count the number of chromosomes in metaphase spreads of the nuclei. Since direct counting under the microscope is tedious and error prone, it is recommended that pictures be taken on which the counted chromosomes can be marked. We use a Zeiss Axiophot microscope equipped with a video camera. Pictures are stored as graphics files and printed out for counting chromosomes. For additional information regarding karyotype analysis, see (63). X-Gal Staining of ES Cells Containing an Inserted and Expressed lacZ Gene In cases where the connexin gene is expressed already in the embryo at the blastocyst stage and in ES cells, it may be useful to monitor the function of a

212

THEIS ET AL.

reporter gene inserted into the locus of the removed connexin gene. While expression of green fluorescent protein (GFP) can be monitored directly by using standard equipment for fluorescence microscopy (64), expression of lacZ is best monitored by staining with X-Gal. Wash cells twice with PBS and fix for 5 min with 0.25% glutaraldehyde in PBS. Do not overfix. Wash three times with PBS to remove fixative and add X-Gal (5-bromo-4-chloro-3-indolyl- ␤ - D -galactopyranoside) staining solution (see Materials). Protect X-Gal staining solution from light. Incubate 2–24 h at 37°C. Make sure that the pH of the staining solution is around 7.3, since this avoids background staining by endogenous acidic ␤-galactosidase. When staining is sufficient, remove staining solution (can be reused several times), wash twice with PBS, and store at 4°C in 70% (v/v) glycerol. The use of a lacZ gene with nuclear localization signal, compared with cytosolic ␤-galactosidase, facilitates discrimination of transgene-derived bacterial ␤-galactosidase from endogenous background activity. For further information see (63). Preparation of HM-1 Cells for Blastocyst Injections Only small numbers of ES cells are needed for blastocyst injection. We routinely work with a subconfluent 25-cm 2 flask. Passage the cells 1 day before injection. On the day of injection change medium early in the morning. Start trypsinization only when blastocysts are available and the microinjection setup is ready. Aspirate medium, and wash once with PBS and once with 2 ml of trypsin solution. Add 3 ml of trypsin and incubate 5–10 min at 37°C to achieve a single-cell suspension (check under microscope and continue incubation with trypsin, if necessary). Add 5 ml of medium and pipet up and down gently. Spin for 2 min at 1200 rpm, aspirate medium, and resuspend cells in 5 ml of injection medium (see Materials). Store cells on ice until they are needed. Inject 8 to 15 cells per blastocyst using an injection setup as described elsewhere (63). The whole blastocyst injection procedure and setting up of a colony of transgenic mice are described extensively in (63, 65). Materials HM-1 culture medium. Glasgow modification of Eagle’s medium (BHK-21 medium/Glasgow MEM), without tryptose phosphate broth, and with L-glutamine (Gibco-BRL 21710-025). Supplements: sodium pyruvate (100 mM, Gibco-BRL 11360-039), nonessential amino acids (100⫻, Gibco-BRL), fetal calf serum, newborn calf serum, 2-mercaptoethanol (0.1 M stock), leukemia inhibitory factor (LIF). Preparation of 2-mercaptoethanol stock solution. Add 0.2 ml of 2-mercaptoethanol to 28.2 ml of distilled water (tissue culture tested, Gibco-BRL) and filter

through 0.2-␮m-pore filter. Store at ⫺20°C. The final concentration in the medium is 10 ⫺4 M. Preparation of complete HM-1 culture medium. Add the following solutions to 500 ml of BHK-21 medium: 5.6 ml sodium pyruvate, 5.6 ml nonessential amino acids, 28 ml fetal calf serum, 28 ml newborn calf serum, 570 ␮l 2-mercaptoethanol, 570 ␮l LIF (this amount will vary from batch to batch) if using selfprepared LIF. For the preparation of LIF, see (66) or follow manufacturers’ recommendations. Store at 4°C and use for up to 1 week. Preparation of gelatin solution. Use swine skin type I gelatin (Sigma G-2500). Prepare a 1% solution in cell culture-grade water. Autoclave, leave overnight, and autoclave again the following day. Store at 4°C. Warm before preparing 0.1% working solution. Trypsin solution. For 500 ml of trypsin solution, add 390 ml PBS [sterile, cell culture grade, calcium and magnesium free (Gibco-BRL)], 100 ml sterile EDTA (1.85 g/liter), 5 ml trypsin solution (Gibco-BRL), and 5 ml chicken serum (Gibco-BRL). Mix and store as 100-ml aliquots at ⫺20°C. Once thawed, keep aliquots at 4°C. 10⫻ HEPES-buffered saline (HBS) for electroporation. Dissolve 16 g NaCl, 0.74 g KCl, 0.252 g Na 2HPO 4, 2 g D-glucose (Dextrose), and 10 g Hepes in 180 ml cell culture-grade water, adjust pH to 7.2, adjust volume to 200 ml, and filter through 0.2-␮m-pore filter. Store at ⫺20°C. To prepare 1⫻ HBS, dilute with sterile water which brings pH to 7.05. Preparation of HAT and HT media. HAT and HT media were bought as 50⫻ stocks from Sigma but used as if they were 100⫻ concentrated. Alternatively, for preparation of 100⫻ HAT solution, dissolve 272.2 mg hypoxanthine in 2 ml 1 N NaOH. Add 8 ml cell culturegrade water and mix well. Add another 10 ml of water and mix until most of the hypoxanthine is dissolved. Add 1 ml of 1 N NaOH and mix until hypoxanthine is dissolved. Add 75 ml of water, then add 48.4 mg thymidine, followed by a further 100 ml of water. Then add 4 ml aminopterin (1 mM) and mix again (aminopterin is poisonous). Filter through 0.2-␮m filter and store as 20-ml aliquots at ⫺20°C. To prepare aminopterin solution (1 mM), suspend 100 mg in 45 ml of cell culture-grade water. Add about 1 ml 1 N NaOH to dissolve aminopterin completely. Sterilize as above and store at ⫺20°C. G418. Prepare at 40 mg/ml (active concentration) in PBS and filter-sterilize. Store at ⫺20°C. Final concentration is 350 – 400 ␮g/ml. Gancyclovir (M r 255.23). Prepare as 1000⫻ stock solution (2 mM). Dissolve 5.1 mg in 10 ml PBS and filter-sterilize. Store in 1.5-ml aliquots at ⫺20°C. 6-Thioguanine (6-TG). Prepare as 1000⫻ stock solution (2 mM). Dissolve 50 mg in 1 ml of 1 N NaOH.

213

TARGETED DELETION AND REPLACEMENT OF CONNEXIN DNA

Add water up to 25 ml, dissolve completely, and filtersterilize. Store as 5-ml aliquots at ⫺20°C. Freezing medium for ES cells. Freezing medium contains complete ES medium plus 10% serum (20% final concentration) and 10% DMSO. To freeze ES cells, prepare 2⫻ freezing medium as follows: 13.3 ml complete ES medium, 4 ml DMSO (ultrapure), 2.7 ml fetal calf serum. Store at ⫺20°C. Lysis buffer. To 100 mM Tris–HCl, pH 8.5, 5 mM EDTA, 0.2% SDS, 200 mM NaCl (store at room temperature), add 200 ␮g/ml proteinase K (20 mg/ml stock in water, store at ⫺20°C) immediately before use. X-Gal staining solution. Combine 0.5 mg/ml X-Gal [stock: 4% in dimethylformamide (DMF), store protected from light at ⫺20°C], 2 mM MgCl 2, 5 mM K 4Fe(CN) 6, and 5 mM K 3Fe(CN) 6 dissolved in PBS. Store at 4°C. Filter immediately prior to use to remove crystals. Injection medium. Combine 40 ml DMEM, 25 mM Hepes (Sigma D-6171), 3.75 ml fetal calf serum (heat inactivated), 0.5 ml glutamine (100⫻), 0.5 ml penicillin/streptomycin (100⫻, Gibco-BRL, stored as 1-ml aliquots at ⫺20°C), and 5.25 ml DMEM (GibcoBRL). Store at 4°C for up to 2 weeks.

WORKING EXPERIENCE AND FUTURE DIRECTIONS Use of PCR for Early Identification of Targeted Clones In our experience, detection of homologous recombination by PCR is superior to Southern blot hybridization for rapid identification of targeted clones, since it saves cost and labor. This can be deduced from Table 1 (experiments 1, 3, 10, and 12–14). Only few PCRpositive clones turned out to be negative when assessed by Southern blot analysis. In experiment 7 (Table 1), 15 of 24 PCR-positive clones have been analyzed by Southern blot hybridization. Twelve of these clones had undergone homologous recombination, which left three false-positive clones as judged by PCR analysis. The yield of useful clones is lower, however, because only 4 of 12 homologously recombined clones showed cointegration of the third loxP site which was located distant with respect to the floxed marker DNA (see Fig. 1d). Experimental results relying on Southern blot analysis are useful to demonstrate the genetic condition of false-positive clones: Experiment 11 in Table 1 shows, after the second gene targeting step of the doublereplacement procedure, that many ES cell clones which

TABLE 1 Efficacy of Homologous Recombination and cre-Mediated Deletion in HM-1 ES Cells a Number of clones (N) Southern blot hybridization No.

Method

Selection

Isolated

PCR positive

Total

Positive

Negative 1 random integration 10 random integration 7 n.d. 58 n.d. 154 random integration 2 wild type 118 n.d. 1 wild type 2 random integration 8 no cointegration 28 no deletion 4 mixed colonies 6 false deletion 51 no deletion 5 false deletion 0 50 reversion to wild type 0 0 0

1 2 3 4 5

HR HR HR HR HR

HAT HAT/Ganc. G418 G418 G418

96 177 200 250 756

17 14 20 n.d. 162

17 14 20 79 162

16 4 13 21 6

6 7

HR/3loxP HR/3loxP

G418/Ganc. G418

233 120

n.d. 49

119 15

1 4

8

cre

G418*

192

n.d.

46

8

9

cre

Ganc.

120

n.d.

57

1

10 11 12 13 14

DR DR DR DR DR

6-TG 6-TG 6-TG 6-TG 6-TG

361 54 156 316 166

136 n.d. 18 34 10

38 54 13 34 10

38 4 13 34 10

a HR, homologous recombination (first/single targeting step); DR, double replacement (second targeting step); cre, transient transfection with cre; Ganc, gancyclovir; 6-TG, 6-thioguanine; n.d., not determined; * replica plating; HAT, HAT selection; 3loxP, construct contains three loxP sites.

214

THEIS ET AL.

had not undergone homologous recombination were wild type, probably due to reversion. This has also been observed by others (14). The first gene targeting step with an HPRT minigene or a neomycin resistance gene often led to resistant colonies with random integration of the targeting vector (see Table 1, experiments 1, 2, 5, and 7). Both observations agree with theoretical considerations: 6-TG resistance cannot be achieved by random integration of a targeting vector, because the necessary loss of the HPRT marker gene in the targeted locus cannot be obtained by random integration. This is different from neomycin or HAT resistance, which can be selected for by insertion of the DNA at the locus to be targeted or elsewhere in the genome. A randomly integrated transgene can lead to a falsepositive result in PCR analysis. The process leading to a PCR product whose template is not contingent but composed of two separate templates that are positioned in trans relative to each other has been described in the literature as splicing by overlap extension PCR (SOE-PCR) (55). It is a convenient method for PCR mutagenesis. The same situation exists in an ES cell clone with a randomly integrated transgene: the binding site for the internal primer may be located on a different chromosome with respect to the target locus that contains the binding site for the external primer. During the first rounds of amplification, the polymerase generates single-stranded DNA molecules which include sequences of one homology arm of the targeting vector. They may serve now as a large primer that binds to the endogenous region of the targeted locus. Elongation of these DNA molecules far beyond the homology region leads to a DNA template that closely resembles the targeted locus. Now both primer binding sites are provided in trans. The external primer as well as the internal primer can bind the newly generated double-stranded DNA molecule and amplify this artificial template which has exactly the same size and base sequence as the template region of the correctly targeted locus. Of course, this kind of false-positive PCR should not be as efficient as true positive PCR (Fig. 1a), because in the latter case the primer binding sites are provided in cis from the beginning of the PCR. SOE-PCR can be used to establish a PCR for detection of homologous recombination (54), but with poor results in our laboratory (Table 1, experiment 5). Relationship between Karyotype and Frequency of Germline Transmission All germline competent clones derived from the HM-1 ES cell line that have been generated to date in our laboratory exhibited at least 74% euploidy. However, a highly euploid karyotype of homologously recombined ES cells is necessary but not sufficient for germline transmission, since blastocyst injections of several clones with high degree of euploidy have pro-

duced mice that were highly chimeric but not germline competent (results not shown). Nevertheless, karyotypes should be determined: after injection of two clones without germline transmission, karyotyping revealed that both clones were 70 –90% aneuploid (41 chromosomes). Aneuploidy is not uncommon among isolated ES cell clones, especially after a second transfection. Therefore, it is most important to isolate a sufficiently large number (around 10) of correctly targeted ES cell clones, since only some of them might be germline competent. Comparison of Cre/loxP Technique and Double Replacement Method for Generation of Floxed Alleles We have carried out two experiments (Nos. 9 and 10 in Table 1) designed to generate the same final recombinant allele which was a floxed connexin-coding region followed by lacZ-coding DNA. In experiment 9, an ES cell clone with three loxP sites generated by homologous recombination was transiently transfected with a Cre-encoding plasmid for partial deletion of the floxed selection marker DNA (Fig. 1d). Only 1 of 57 resistant colonies displayed the desired partial deletion. In experiment 10, the double-replacement method was used to replace an HPRT minigene residing in the connexin locus with a floxed connexin-coding region followed by lacZ-coding DNA (Fig. 1b). Thirty-eight percent of the resistant colonies had undergone homologous recombination as assessed by PCR and all PCRpositive clones, after further analysis by Southern blot hybridization, were confirmed to be correctly recombined. Thus, the double-replacement technique led to a more than 20-fold yield compared with the Cre/loxP method. According to these results, the doublereplacement technique appears superior to transient transfection with a Cre-encoding plasmid for the generation of a floxed allele. This conclusion is based on theoretical considerations as well as our own experience using both techniques: First, when using a targeting vector with three loxP sites, there is one loxP site located in the homology region, since the floxed part of the gene is part of the endogenous targeted sequence (Fig. 1d). Lack of cointegration of the distant loxP site has repeatedly been reported (67– 69) and occurred in experiment 7 (Table 1). Thus, many of the expanded PCR-positive clones were useless as deduced from later Southern blot analysis. This problem did not arise when we performed two sequential homologous recombinations. After the first targeting event, the region to be flanked was deleted from the recombinant allele. During the second targeting event, the whole floxed region of the targeting vector had no counterpart on the recombined allele, hence no recombination could occur between the two loxP sites (Figs. 1b, 2). Each PCR-positive recombinant clone should in theory be correctly targeted. This has been observed using the double-replacement method

TARGETED DELETION AND REPLACEMENT OF CONNEXIN DNA

(second homologous recombination) for generation of a floxed allele (Table 1, experiment 10). Second, double replacement circumvents the need for transient transfection to delete the selection marker cassette which proved to be quite inefficient compared with the introduction of the floxed gene by a second step of homologous recombination (Table 1, experiments 9 and 10). If, for several reasons, a conditional gene deletion becomes necessary to elucidate gene function, two strategies can be used: the Cre/loxP system and the double-replacement strategy (see Figs. 1b, 1d). Both will allow systemic as well as conditional gene ablation. One can use the Cre/loxP technique to generate an allele with three loxP sites in the first step via homologous recombination and to isolate clones with the floxed allele in the second step via transient transfection with a Cre-encoding plasmid (Fig. 1d). ES cell clones with total deletion can be used for generation of the conventional null mutant mouse. However, the latter ES cell clones have undergone two sequential electroporations. The double-replacement strategy requires only one electroporation step for the generation of conventional null mutants compared with the Cre/ loxP method. Of course, it is always possible to directly inject ES cell clones into blastocysts after the first electroporation step, resulting in an allele with three loxP sites. Breeding with a deleter mouse strain will subsequentially lead to the conventional null mutant mouse line, avoiding the second electroporation step. Furthermore, mice harboring an allele with three loxP sites have been repeatedly described as conditionally deficient mice (69, 70). However, potential side effects, due to the persisting selection marker DNA, are possible (71, 72). Therefore, the generation of mice carrying a floxed gene without selection marker DNA is preferable. Again, the double-replacement strategy appears to be more advantageous as it has yielded a larger number of correctly targeted clones for generation of conditional deficient mouse mutants. There is no cointegration problem of loxP sites in the first step (compare Fig. 1b with Fig. 1d) and a higher frequency of useful clones in the second step of the doublereplacement method compared with the Cre/loxP technique (see Table 1, compare experiments 9 and 10). The use of an HPRT minigene for conventional gene targeting always allows further use of the germlinecompetent ES cell clone to be targeted once again to introduce flanking loxP sites near the gene of interest. In addition, targeted replacement or conditional replacement can be performed based on this ES cell clone. One caveat regarding the use of an HPRT minigene for the generation of a null mutant is a possible effect of HPRT overexpression in certain cell types. However, transgenic mice have been generated that overexpress HPRT but do not show any obvious phe-

215

notypic abnormality (73). To avoid the minimal risk of unwanted side effects resulting from overexpression of HPRT, one can flank the HPRT minigene with loxP sites and delete the HPRT minigene in the mouse by crossing with a deleter mouse strain, leaving only a single loxP site at the targeted locus. Such a plasmid is available on request ([email protected]). For generating gene replacements or subtle mutations, the Cre/loxP technique is superior to the doublereplacement method, since it allows omission of the second round of transfection, which can reduce germline transmission. In this case transgenic animals, heterozygous for the subtle mutation and harboring the floxed selection marker DNA, can be crossed to a mouse strain that expresses cre ubiquitously or early in development (Fig. 1c). After one mouse generation, a new mutant mouse strain can be generated with a subtle mutation and only a few additional base pairs due to a single remaining loxP site. However, if gene replacement or subtle mutation leads to early lethality, it might be necessary to perform conditional genetic manipulation. Compensation among Connexins: High-Resolution Analysis To analyze functional compensation among connexins, it is necessary to generate double or triple connexin-deficient mice (74). One or more of these connexin-deficient mutations will have to be conditional. In addition to connexin loci, other transgenic loci for cre expression are involved. The number of loci should be kept as small as possible to save time and space for the mouse crosses. One way to keep the number of loci involved in the crossings small is to insert an inducible Cre/hormone receptor fusion protein gene into the target locus (Fig. 2). Mice harboring a floxed allele of a given connexin locus can be mated with mice harboring DNA coding for an inducible recombinase at the second copy of the same locus. Each cell expressing the floxed connexin is also expressing the posttranslationally inducible cre/hormone receptor fusion gene. On induction, functionally ubiquitous deletion of the particular connexin can take place. Unexpected functions of the targeted connexin can be revealed, avoiding possible embryonic or perinatal lethality. The second advantage, especially for studies of functional complementation in conditional double or triple connexin mutant mice, is that the cre/hormone receptor fusion gene is expressed under control of the connexin promoter. Therefore, the insertion of a posttranslationally inducible cre gene allows a very detailed analysis of compensation among connexins. On induction, the floxed connexin allele which shares the locus with the inducible cre fusion gene is deleted in each cell where it is expressed. Every other connexin DNA that is floxed is also deleted, but only in cells where both

216

THEIS ET AL.

connexins are coexpressed. Hence, compensatory activity of this second connexin can be excluded. The second connexin with a possible compensatory role might have other functional roles as well. If only the coexpressed connexin is deleted, these roles can be fulfilled, and no secondary effects may disturb the compensation study. It would be interesting to compare these conditional double deficient mice with constitutive double deficient animals. With this precision at the cellular level, it might be possible to determine whether or not compensatory effects are cell autonomous. Using this allelic configuration, one should keep in mind that due to replacement of the connexin-coding region by the cre gene, the targeted connexin gene exists in the heterozygous state, because the inducible recombinase allele is also a null allele. If the connexin deletion has subtle effects in the heterozygous mouse mutant, this has to be taken into account (75). To circumvent these effects, bicistronic expression of an inducible cre/hormone receptor fusion gene similar to (76) could be an alternative.

b. Nuclear Localization of Reporter Proteins Both GFP and ␤-galactosidase are located in the cytoplasm. Thus, nuclear localization of ␤-galactosidase (32) and GFP could be used for analysis. It might be advantageous in brain tissue for easier identification of certain neurons. c. Mitochondrial Localization Nuclear and mitochondrial targeting of GFP has been shown previously (81). Mitochondrial targeting leads to higher sensitivity, and endogenous fluorescence in processed mouse tissue can be distinguished more easily from nuclear or mitochondrial GFP fluorescence than from cytoplasmic GFP fluorescence (81). d. Different Epitope Tags A manipulated lacZ gene that codes for an additional epitope tag could be used. If changes in expression patterns are suspected in double or triple connexin gene-deficient mice, staining with antibodies should allow distinction between different sources of ␤-galactosidase activity.

Use of Different Reporter Genes to Monitor Different Connexin Deletions Reporter genes incorporated into the targeting vector serve to indicate transcriptional activity of the corresponding connexin promoters in the homozygous deficient state (15, 32). In the heterozygous state, which phenotypically often resembles the wild-type situation, this analytical approach is far less expensive, faster, and easier compared with in situ hybridization. One caveat is that no other information besides transcriptional activity can be deduced. The half-life of connexins in several cell types is relatively short (77, 78), whereas variants of green fluorescent protein and ␤-galactosidase are very stable proteins. Nevertheless, reporter genes can be used to screen mutual effects of connexins on the transcriptional level, for example, in mutant mice deficient for two connexin genes. It is possible that loss of one connexin alters the expression of another connexin (31). To detect changes in expression of one connexin caused by the loss of another connexin, it might be preferable to use different reporter genes or at least differently localized reporter proteins. There are several ways to cope with this problem: a. Different Reporter Genes Mice showing high expression of GFP are viable, demonstrating low or zero toxicity (79). The great advantage of GFP is that the examination of living tissue is possible (79). However, in fixed or processed mouse tissue, high endogenous fluorescence, overlapping with the emission spectrum of GFP, has been reported (80).

ACKNOWLEDGMENTS We are grateful to Dominik Eckardt, Gaby Hallas, Olaf Krueger, Karen Maass, Thomas Ott, Barbara Teubner, and Eva Thoenissen for providing unpublished results summarized in Table 1. Martin Theis received a stipend from the Graduiertenkolleg “Pathogenese von Krankheiten des Nervensystems.” Work in our laboratory was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 284/C1, Z2, and SFB400/B3), the Deutsche Krebshilfe, DEBRA (UK), the Biomed program of the EU, and the Fonds der Chemischen Industrie.

REFERENCES 1. Goodenough, D. A., Goliger, J. A., and Paul, D. L. (1996) Annu. Rev. Biochem. 65, 475–502. 2. Condorelli, D. F., Parenti, R., Spinella, F., Salinaro, A. T., Belluardo, N., Cardile, V., and Cicirata, F. (1998) Eur. J. Neurosci. 10, 1202–1208. 3. Soehl, G., Degen, J., Teubner, B., and Willecke, K. (1998) FEBS Lett. 428, 27–31. 4. Manthey, D., Bukauskas, F., Lee, C. G., Kozak, C., and Willecke, K. (1999) J. Biol. Chem., in press. 5. Kumar, N. M., and Gilula, N. B. (1996) Cell 84, 381–388. 6. White, T. W., and Paul, D. L. (1999) Annu. Rev. Physiol. 61, 283–310. 7. Temme, A., Buchmann, A., Gabriel, H. D., Nelles, E., Schwarz, M., and Willecke, K. (1997) Curr. Biol. 7, 713–716. 8. Gabriel, H. D., Jung, D., Buetzler, C., Temme, A., Traub, O., Winterhager, E., and Willecke, K. (1998) J. Cell Biol. 140, 1453– 1461. 9. Gros, D. B., and Jongsma, H. J. (1996) Bioessays 18, 719 –730. 10. Bruzzone, R., and Ressot, C. (1997) Eur. J. Neurosci. 9, 1– 6.

TARGETED DELETION AND REPLACEMENT OF CONNEXIN DNA 11. Simon, A. M., and Goodenough, D. A. (1998) Trends Cell Biol. 8, 477– 483. 12. Willecke, K., Kirchhoff, S., Plum, A., Temme, A., Thoenissen, E., and Ott, T. (1999) in Gap Junction-Mediated Intercellular Signalling in Health and Disease (Cardew, G., Ed.), pp. 76 –96, Wiley, Chichester. 13. Thomas, K. R., and Capecchi, M. R. (1987) Cell 51, 503–512. 14. Gschwind, M., and Huber, G. (1998) Mol. Cell. Biol. 18, 4651– 4658. 15. Plum, A., Rosentreter, B., Traub, O., and Willecke, K. (1998) Eur. J. Cell Biol. Suppl. 48 (Vol.75), 23. 16. Sauer, B. (1998) Methods Companion Methods Enzymol. 14, 381–392. 17. Sauer, B. (1993) Methods Enzymol. 225, 890 –900. 18. Sauer, B. (1987) Mol. Cell. Biol. 7, 2087–2096. 19. Sauer, B., and Henderson, N. (1989) Nucleic Acids Res. 17, 147–161. 20. Gu, H., Marth, J. D., Orban, P. C., Mossmann, H., and Rajewsky, K. (1994) Science 265, 103–106. 21. DiSanto, J. P., Mueller, W., Guy-Grand, D., Fischer, A., and Rajewsky, K. (1995) Proc. Natl. Acad. Sci. USA 92, 377–381. 22. Lakso, M., Sauer, B., Mosinger, B., Lee, E. J., Manning, R. W., Yu, S. H., Mulder, K. L., and Westphal, H. (1992) Proc. Natl. Acad. Sci. USA 89, 6232– 6236. 23. Agah, R., Frenkel, P. A., French, B. A., Michael, L. H., Overbeek, P. A., and Schneider, M. D. (1997) J. Clin. Invest. 100, 169 –179. 24. Kuehn, R., Schwenk, F., Aguet, M., and Rajewsky, K. (1995) Science 269, 1427–1429. 25. St-Onge, L., Furth, P. A., and Gruss, P. (1996) Nucleic Acids Res. 24, 3875–3877. 26. Danielian, P. S., Muccino, D., Rowitch, D. H., Michael, S. K., and McMahon, A. P. (1998) Curr. Biol. 8, 1323–1326. 27. Brocard, J., Warot, X., Wendling, O., Messaddeq, N., Vonesch, J. L., Chambon, P., and Metzger, D. (1997) Proc. Natl. Acad. Sci. USA 94, 14559 –14563. 28. Schwenk, F., Kuehn, R., Angrand, P. O., Rajewsky, K., and Stewart, A. F. (1998) Nucleic Acids Res. 26, 1427–1432. 29. Kellendonk, C., Tronche, F., Casanova, E., Anlag, K., Opherk, C., and Schuetz, G. (1999) J. Mol. Biol. 285, 175–182. 30. Reaume, A. G., de Sousa, P. A., Kulkarni, S., Langille, B. L., Zhu, D., Davies, T. C., Juneja, S. C., Kidder, G. M., and Rossant, J. (1995) Science 267, 1831–1834. 31. Nelles, E., Buetzler, C., Jung, D., Temme, A., Gabriel, H. D., Dahl, U., Traub, O., Stuempel, F., Jungermann, K., Zielasek, J., Toyka, K. V., Dermietzel, R., and Willecke, K. (1996) Proc. Natl. Acad. Sci. USA 93, 9565–9570. 32. Gong, X., Li, E, Klier, G., Huang, Q., Wu, Y., Lei, H., Kumar, N. M., Horwitz, J., and Gilula, N. B. (1997) Cell 91, 833– 843. 33. White, T. W., Goodenough, D. A., and Paul, D. L. (1998) J. Cell Biol. 143, 815– 825. 34. Elfgang, C., Eckert, R., Lichtenberg-Frate, H., Butterweck, A., Traub, O., Klein, R. A., Huelser, D. F., and Willecke, K. (1995) J. Cell Biol. 129, 805– 817. 35. Bruzzone, R., White, T. W., and Goodenough, D. A. (1996) Bioessays 18, 709 –718. 36. Stacey, A., Schnieke, A., McWhir, J., Cooper, J., Colman, A., and Melton, D. W. (1994) Mol. Cell. Biol. 14, 1009 –1016. 37. Moore, R. C., Hope, J., McBride, P. A., McConnell, I., Selfridge, J., Melton, D. W., and Manson, J. C. (1998) Nat. Genet. 18, 118 –125. 38. Askew, G. R., Doetschman, T., and Lingrel, J. B. (1993) Mol. Cell. Biol. 13, 4115– 4124.

217

39. Wu, H., Liu, X., and Jaenisch, R. (1994) Proc. Natl. Acad. Sci. USA 91, 2819 –2823. 40. Magin, T. M., McWhir, J., and Melton, D. W. (1992) Nucleic Acids Res. 20, 3795–3796. 41. Hormuzdi, S. G., Penttinen, R., Jaenisch, R., and Bornstein, P. (1998) Mol. Cell. Biol. 18, 3368 –3375. 42. Detloff, P. J., Lewis, J., John, S. W. M., Shehee, W. R., Langenbach, R., Maeda, N., and Smithies, O. (1994) Mol. Cell. Biol. 14, 6936 – 6943. 43. Yang, B., Kirby, S., Lewis, J., Detloff, P. J., Maeda, N., and Smithies, O. (1995) Proc. Natl. Acad. Sci. USA 92, 11608 –11612. 44. Lewis, J., Yang, B., Detloff, P., and Smithies, O. (1996) J. Clin. Invest. 98, S3–S5. 45. van der Neut, R. (1997) J. Neurosci. Methods 71, 19 –27. 46. Kuehn, R., and Schwenk, F. (1997) Curr. Opin. Immunol. 9, 183–188. 47. Torres, R. M., and Kuehn, R. (1997) Laboratory Protocols for Conditional Gene Targeting, Oxford Univ. Press, Oxford. 48. Akagi, K., Sandig, V., Vooijs, M., Van der Valk, V. M., Giovannini, M., Strauss, M., and Berns, A. (1997) Nucleic Acids Res. 25, 1766 –1773. 49. Davies, T. C., Barr, K. J., Jones, D. H., Zhu, D., and Kidder, G. M. (1996) Dev. Genet. 18, 234 –243. 50. De Sousa, P. A., Juneja, S. C., Caveney, S., Houghton, F. D., Davies, T. C., Reaume, A. G., Rossant, J., and Kidder, G. M. (1997) J. Cell Sci. 110, 1751–1758. 51. Oyamada, Y., Komatsu, K., Kimura, H., Mori, M., and Oyamada, M. (1996) Exp. Cell Res. 229, 318 –326. 52. Oyamada, M., Oyamada, Y., Komatsu, K., Takamatsu, T., and Mori, M. (1998) in Gap Junctions (Werner, R., Ed.), pp. 352–356, IOS Press, Amsterdam. 53. http://jaxmice.jax.org/html/nomenclature/nomen_129.shtml (1998). 54. Koentgen, F., and Stewart, C. L. (1993) Methods Enzymol. 225, 878 – 890. 55. Dieffenbach, C. W., and Dveksler, G. S. (1995) PCR Primer: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Plainview, NY. 56. O’Gorman, S., Dagenais, N. A., Qian, M., and Marchuk, Y. (1997) Proc. Natl. Acad. Sci. USA 94, 14602–14607. 57. Wang, F., Nemes, A., Mendelsohn, M., and Axel, R. (1998) Cell 93, 47– 60. 58. Schwenk, F., Baron, U., and Rajewsky, K. (1995) Nucleic Acids Res. 23, 5080 –5081. 59. Lakso, M., Pichel, J. G., Gorman, J. R., Sauer, B., Okamoto, Y., Lee, E., Alt, F. W., and Westphal, H. (1996) Proc. Natl. Acad. Sci. USA 93, 5860 –5865. 60. Lallemand, Y., Luria, V., Haffner-Krausz, R., and Lonai, P. (1998) Transgenic Res. 7, 105–112. 61. Laird, P. W., Zijderveld, A., Linders, K., Rudnicki, M. A., Jaenisch, R., and Berns, A. (1991) Nucleic Acids Res. 19, 4293. 62. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Plainview, NY. 63. Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994) Manipulating the Mouse Embryo: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Plainview, NY. 64. Clontech (1997) Living Colours User Manual (PT2040-1), Clontech Laboratories, Palo Alto, CA. 65. Stewart, C. L. (1993) Methods Enzymol. 225, 823– 855. 66. Smith, A. G. (1991) J. Tissue Culture Methods 3, 89 –94. 67. Marth, J. D. (1996) J. Clin. Invest. 98, S47–S50.

218

THEIS ET AL.

68. Priatel, J. J., Sarkar, M., Schachter, H., and Marth, J. D. (1997) Glycobiology 7, 45–56. 69. Tarutani, M., Itami, S., Okabe, M., Ikawa, M., Tezuka, T., Yoshikawa, K., Kinoshita, T., and Takeda, J. (1997) Proc. Natl. Acad. Sci. USA 94, 7400 –7405. 70. Yu, R. N., Ito, M., Saunders, T. L., Camper, S. A., and Jameson, J. L. (1998) Nature Genet. 20, 353–357. 71. Meyers, E. N., Lewandoski, M., and Martin, G. R. (1998) Nat. Genet. 18, 136 –141. 72. Nagy, A., Moens, C., Ivanyi, E., Pawling, J., Gertsenstein, M., Hadjantonakis, A. K., Pirity, M., and Rossant, J. (1998) Curr. Biol. 8, 661– 664. 73. Stout, J. T., Chen, H. Y., Brennand, J., Caskey, C. T., and Brinster, R. L. (1985) Nature 317, 250 –252. 74. Houghton, F. D., Thoenissen, E., Kidder, G. M., Naus,

75.

76. 77. 78. 79. 80. 81.

C. C., Willecke, K., and Winterhager, E. (1999) Dev. Genet. 24, 5–12. Guerrero, P. A., Schuessler, R. B., Davis, L. M., Beyer, E. C., Johnson, C. M., Yamada, K. A., and Saffitz, J. E. (1997) J. Clin. Invest. 99, 1991–1998. Lindeberg, J., and Ebendal, T. (1999) Nucleic Acids Res. 27, 1552–1554. Laird, D. W. (1996) J. Bioenerg. Biomembr. 28, 311–318. Beardslee, M. A., Laing, J. G., Beyer, E. C., and Saffitz, J. E. (1998) Circ. Res. 83, 629 – 635. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., and Nishimune, Y. (1997) FEBS Lett. 407, 313–319. Zhuo, L., Sun, B., Zhang, C. L., Fine, A., Chiu, S. Y., and Messing, A. (1997) Dev. Biol. 187, 36 – 42. Tsien, R. Y. (1998) Annu. Rev. Biochem. 67, 509 –544.