ABSTRACT
Repair of DNA damage resulting in double-strand breaks (DSBs) is controlled by gene products executing homologous recombination or end-joining pathways. The MRE11 gene has previously been implicated in DSB repair in the yeast Saccharomyces cerevisiae. Here we have developed a methodology to study the roles of the murine Mre11 homolog in pluripotent embryonic stem cells. Using a gene targeting approach, a triple LoxP site cassette was inserted into a region of MRE11 genomic DNA flanking conserved phosphodiesterase motifs. The addition of Cre recombinase activity promotes deletions of three types that can be scored. We find that deletion at phosphodiesterase motif III encoded in the N-terminus of Mre11 is acheived in the presence of a wild-type MRE11 allele. However, when the wild-type MRE11 allele is inactivated by gene targeted insertion of a neo marker, only Cre recombination events that allow expression of wild-type Mre11 protein are observed. Therefore, Mre11 is required for normal cell proliferation. This methodology introduces a means to study important regions of essential genes in cell culture models.
Repair of chromosome damage in eukaryotes is dictated by a group of repair pathways and dependent on the type of chromosome break or DNA modification. For double-stranded breaks (DSBs) there are two principle repair mechanisms that have been identified. DSBs in the yeast Saccharomyces cerevisiae are mainly repaired by homologous recombination, where a large group of genes have been demonstrated to be utilized (RAD50-RAD58, MRE2, MRE11 and XRS2; reviewed in 1 ,2 ). A second pathway in eukaryotes involves the joining of DSBs with little requirement for DNA homology (3 ). These end-joining pathways have been adopted in developmentally programed gene rearrangements, such as V(D)J recombination. A subset of the genes involved in homologous recombination may also be utilized in end-joining pathways (4 -6 ).
Mutation of the S.cerevisiae MRE11 gene blocks meiotic recombination, has a hyper-recombinational phenotype and generates sensitivity to DSB damage in mitotic yeast (7 ). Mutation of the Schizosaccharomyces pombe MRE11 homolog (rad32) also causes increased chromosome loss (8 ). Two other yeast genes, RAD50 and XRS2, share the same phenotypes as MRE11 (9 ,10 ). Mre11 and Rad50 have been shown to interact using yeast two-hybrid analysis (11 ) and co-immunoprecipitation from human cells (12 ). Rad50 is a DNA and nucleotide binding protein of unknown function (13 ).
MRE11 and RAD50 are related to bacterial genes involved in chromosome stability in Escherichia coli (14 ). The products of sbcC and sbcD control destabilization of long cruciform or palindromic DNA sequences in the E.coli genome that may interfere with DNA synthesis (15 ). sbcC/D are also similar to nucleases of the UvrA family (16 ) and have an ATP-dependent exonuclease activity (17 ). For Mre11, the homology with sbcD is much greater than with other members of the phosphodiesterase superfamily and four regions of common residues are found (motifs I-IV) over ~100 amino acids (1 4 ). These strucutural motifs may indicate that MRE11 also encodes a nuclease.
Mammalian cells also have the capacity to undergo homologous recombination reactions (18 ). The ability to study the requirements for genes in this process has been hampered by the severity of the engineered mutations. For example, mutations in the recA-related DSB repair gene RAD51 are inviable in cell culture and formation of rad51-/- mouse embryos is blocked by the 4- to 8-cell stage (19 ).
Here we have developed a methodology to examine the requirement for MRE11 functions in mammalian cells. We find that inactivation of MRE11 by gene targeted insertion can be accomplished in a conditional manner using the Cre/loxP site-specific recombination system (20 ,21 ). Following addition of Cre to targeted cells, the rescue of loxP deletions in the MRE11 locus is dependent on the status of MRE11 on the other allele. Thus, Cre/loxP recombination has been used to prove an essential requirement for MRE11 in mammalian cell proliferation.
A murine [lambda] phage genomic DNA library, prepared from mouse strain 129 (Clontech), was screened by standard hybridization methodologies with an N-terminal human MRE11 cDNA fragment (1.3 kb) isolated previously (22 ,23 ). Genomic DNA inserts from three phage (18-1, 2-6 and 20-4) were found to have overlapping MRE11 genomic regions. Using an EcoRI-BamHI probe specific for the 5' 720 bp of the human MRE11 cDNA, these three phage were ordered and a restriction enzyme map showing the presence of MRE11 exons identified as shown in Figure 1 . A 1.0 kb HindIII-XhoI fragment and a 0.6 kb HindIII-EcoRI fragment were subcloned into pSKII and resulting inserts analyzed by automated DNA sequencing. Exons corresponding to a 5' region of the human MRE11 cDNA were identified.
A gene targeting vector was constructed by subcloning MRE11 genomic fragments into pPGKneo as follows. The plasmid pPGKneo (24 ), containing the mutant version of the G418 resistance gene expressed from the mouse phosphoglycerate kinase (PGK) promoter, was digested with HindIII and XhoI and blunted by Klenow polymerase extension. A XhoI-EcoRI 5 kb MRE11 genomic fragment isolated from phage 18-1 was subcloned into this site. Second, a 7.5 kb MRE11 EcoRV-NotI genomic fragment isolated from phage 20-4 was then subcloned into a unique SmaI site of this plasmid, resulting in pYX11. MRE11 gene targeting was accomplished by linearization of 50 [mu]g pYX11 at the unique NotI site, followed by ethanol precipitation of the DNA and resuspension in 25 [mu]l sterile H2O.
Murine embryonic stem cells (D3, 129 strain) were cultured as previously described in ES cell medium (DMEM, 20% inactivated fetal calf serum + LIF) (25 ). For introduction of plasmid DNA, 1 * 107 cells were harvested and resuspended in 0.8 ml DME in a 0.4 cm electroporation cuvette (BioRad), 25 [mu]l linearized DNA were added and electroporation conducted in a BioRad Gene Pulser at 0.23 kV, 500 [mu]F. Cells were subsequently plated on a single gelatinized plate in medium overnight. The medium was then replaced with ES medium plus 200 [mu]g/ml G418 for neo selection. Cells were fed with this medium for 7-10 days prior to picking of colonies and/or further analysis. For gene targeting with puromycin resistance plasmids, identical electroporation conditions were used and selection was conducted in ES medium plus 1 [mu]g/ml puromycin.
Genomic DNA was isolated from ES cells by previous methods (22 ,25 ). For analysis of pYX11 gene targeting events at the MRE11 locus, 15 [mu]g genomic DNA was digested with either HindIII or EcoRI for >4 h and fractionated on 0.8% agarose gel electrophoresis. Similarly, for pYX25 gene targeting events genomic DNAs were digested with the same enzymes. Genomic Southern blots were prepared by standard methods (22 ). A phage 18-1 XhoI fragment (1 kb) was subcloned into pSKII, yielding pYX1. Probe 1 was a 0.4 kb XhoI-NcoI fragment from pYX1. Likewise, a phage 2-6 XbaI-EcoRV fragment (1 kb) was subcloned into pSKII, giving pYX9. Probe 2 was a 1 kb XbaI-EcoRV fragment from pYX9. Probe 3 was a 0.5 kb EcoRV-HindIII fragment from pYX7.
The human MRE11 cDNA (23 ) was digested with PstI and EcoRI and subcloned into pET15b. For bacterial expression of an ~70 kDa truncated Mre11 protein, BL21 E.coli cells were grown overnight on an LB plus 100 [mu]g/ml ampicillin plate and a single colony used to innoculate 25 ml LB plus 100 [mu]g/ml carbenicillin that was grown to an optical density of 0.6 OD595, adjusted to 1 mM IPTG and grown for an additional 4 h at 30oC. Cell pellets were harvested by centrifugation and lysates were prepared by sonication in 20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 5 mM imidazole, 6 M urea. Induction of Mre11 was examined by fractionation of lysates with or without IPTG addition by SDS-PAGE, followed by staining with Coomassie brilliant blue. Mre11 protein was first bound and eluted by Ni2+ affinity chromatography (Novagen) according to the supplier's instructions and was fractionated preparatively by SDS-PAGE. The 70 kDa protein was then visualized by KCl precipitation staining (22 ) and a gel fragment removed. The gel fragment containing Mre11 protein was directly used as an immunogen in New Zealand White rabbits (Hazelton Research Products).
For immunoblotting with Mre11 antibody, cell lysates were formed by direct resuspension of cell pellets (1 * 106 cells) in 100 [mu]l 1* SB (50 mM Tris-HCl, pH 7.0, 2% SDS, 360 mM 2-mercaptoethanol, 2.5 mg/ml bromophenol blue), boiled for 5 min, sheared with a 25 gauge needle and fractionated by 7.5 or 12% SDS-PAGE. The SDS-PAGE fractions were transferred to PVDF membranes (NEN) using a BioRad Semi-Dry Electrophoretic Transfer apparatus according to the manufacturer's directions at 20 V for 45 min in 48 mM Tris-HCl, 39 mM glycine, 1.3 mM SDS, 20% methanol, pH 9.2. Blots were blocked in PBS plus 10% Carnation non-fat dry milk for 30 min to overnight at room temperature. Polyclonal anti-Mre11 antiserum was used at a 1:2500 dilution in PBS for 60 min at room temperature. Blots were then washed five times in PBS plus 0.2% Tween-20. Blots were then incubated with 1:5000 protein A/G-HRP (Pierce) for 30-40 min in PBS plus 1% BSA and extensively washed in PBS plus 0.2% Tween-20. Next, blots were developed using a Chemiluminescence Renaissance Kit (NEN) according to the supplier's instructions. Signals were detected on Kodak XR1 film.
For preparation of a loxP-containing gene targeting vector, plasmid pYX11 was modified by the introduction of loxP recognition sites (34 bp) synthesized with two complementary oligonucleotides (AAGCATAACTTCGTATAATGTATGCTATACGAAGTTAT and AAGCATAACTTCGTATAGCATACATTATACGAAGTTAT) containing HindIII restriction sites at their 5'-ends. Following annealing of these two oligonucleotides, the double-stranded oligonucleotides were then subcloned into PYX11 at a HindIII site and the insertion and orientation of single loxP sites identified by DNA sequencing of clones. In addition, plasmid pPGKpuro, containing the puromycin resistance gene expressed from the PGK promoter, was modified so that loxP recognition sites were introduced at each side of the puro gene. This was accomplished by subcloning the PGKpuro fragment (digested with SalI and blunted) into pBS246 (LoxP-containing vector; Gibco BRL) at the HindIII site. A NotI LoxP-puro-LoxP fragment was then subcloned into pYX23 at the EcoRV site, creating pYX25. Gene targeting with pYX25 and ES cell selection was conducted as described above. The pYX25 transfection results in clones with either the wild-type MRE11 allele or the mre11neo allele targeted by the loxP insert.
For Cre recombinase-stimulated deletion at loxP sites in ES cells, the following methodology was adapted. Approximately 1 * 107 309 derivative cells were electroporated with 50 [mu]g pBS185 (Cre recombinase expression vector; Gibco BRL). 309 derivative cells were identified by the pYX25 transfections above and are described in Results. Cells were replated on ES medium without selection at dilutions of 10-3, 10-4 and 10-5 and incubated for 7 days for harvesting of single colonies. ES cell colonies were picked by manipulation with a P200 pipettor, trypsinized and replated in a single 1 cm well of a 24-well plate in ES medium. After expansion of the cells for 4-7 days, cells were frozen in liquid N2 for storage and harvested for genomic DNA analysis.
Alternatively, following electroporation with pBS185 (Cre), cells were plated at different densities after transfection. Aliquots of the transfected cells were plated at various dilutions and harvested at either 2 or 10 days post-transfection. Approximately 1-2 * 106 cells were used for preparation of genomic DNA (1 mg/ml).
ES genomic DNAs (1 mg/ml) were used in PCR and Southern blot analysis to determine loxP-mediated Cre recombination. PCR in the MRE11 genomic region was conducted using two oligonucleotides (YX18, CCCACTGGTGTTTCAAGGGTTGA, and YX19, GGCCAGTACTAGTGAACCTGTTC) and a PCR kit (Expand Long Template PCR System; Boehringer-Mannheim) according to the supplier's instructions. YX18 is located in MRE11 genomic sequence and YX19 is part of pBS246 adjacent to a LoxP site. PCR products were fractionated on 1% agarose gel electrophoresis, transferred to nitrocellulose and blotted with specific hybridization probes in the region. Within this PCR fragment, probe 4 is a 0.6 kb EcoRI-HindIII fragment and probe 2 is as described above.
Yeast mre11 strains are deficient in DSB repair and recombination (7 ). To understand the roles of mammalian MRE11 in DSB repair, we isolated genomic regions of the murine MRE11 locus using our previously identified human MRE11 cDNA (Materials and Methods). The 5'-end of human MRE11 cDNA was used to screen a [lambda] phage 129 mouse strain genomic DNA library and three phage were isolated. Mapping, subcloning and DNA sequencing was then used to identify MRE11 exons in the 5'-end of a putative murine MRE11 open reading frame (ORF). We identified two exons by this approach, corresponding to the interval from amino acid 44 to 119 of the 708 amino acid human MRE11 gene (Fig. 1 A).
MRE11 is a member of a gene family encoding phosphodiesterase functions, such as nucleases (14 ). The N-terminal region of human and S.cerevisiae Mre11, including the 44-119 amino acid region, has greatest similarity with sbcD. By DNA sequencing of the above MRE11 exons within the murine genomic DNA fragments, we found that portions of the phosphodiesterase motifs were present in the mouse ORF. In particular, the second phosphodiesterase motif (II) was partially included in an exon homologous to human Mre11 amino acids 44-69 (II, Fig. 1 B). Within this exon, the murine and human genes were identical. The most conserved portion of motif II among the phosphodiesterase family (DxxILxxGDL) was found (Fig. 1 B). Likewise, an additional exon encoding the third phosphodiesterase motif (III) was found, corresponding to human Mre11 amino acids 86-112. Within the motif III exon, the murine and human genes were identical and the conserved residues PxxxIxGNHD were observed (Fig. 1 B).
In order to examine the role of MRE11 in mammalian cells, we modified the genomic DNA in the motif III region in preparation for gene targeting. Approximately 12.5 kb of mouse MRE11 genomic DNA was subcloned into a gene targeting vector containing a mutant G418 resistance gene (neo) expressed from the phosphoglycerate kinase promoter to make pYX11 (Fig. 2 A). pYX11 was linearized by NotI digestion and transfected into the murine embryonic stem (ES) cell line D3 by electroporation (Materials and Methods). After 1 day, G418 selection was initiated and continued for 1 week according to previous protocols (18 ). G418r colonies were expanded and analyzed by Southern blots for the pattern of integration at the MRE11 locus.
Here we show that the mammalian MRE11 gene is required for normal cell proliferation in a conditional knock-out system we have developed using murine embryonic stem (ES) cells and Cre/loxP. An advantage of our methodology is the ability to score requirements for genes from the pluripotent ES cells without resorting to labor-intensive mouse injections, screening and crossing. The method relies on the differential removal of MRE11-specific exons under conditions of Cre recombinase expression and dependent on the MRE11 genotype at the other allele. Because the Cre recombination outcome is dependent on Mre11, our strategy introduces a methodology to further examine the structural requirements for Mre11. For example, expression of MRE11 cDNA in conjunction with Cre would be anticipated to rescue 1/3 deletions by reconstituting the cell proliferation defect. In addition, screens for other genes overcoming the dependence on MRE11 can be envisaged.
Many other genes have been observed to have early embryonic lethality in mouse models and there are numerous circumstances where ES cells cannot be formed with homozygous mutations in a gene. Within the DSB repair homologous recombination epistasis group, mouse homologs for several of the genes have recently been derived. For murine RAD51, the strand exchange and recA-related gene, double knock-out cells and rad51-/- mice cannot be isolated (19 ). Likewise, the tumor suppressor protein Brca1, which interacts with Rad51 in vivo and in vitro (26 ), has a similar embryonic lethality (27 ). On the other hand, for murine or chicken RAD54, doubly homozygous mice and cell lines can be generated, even though these cells are defective in IR repair and homologous recombination (28 ,29 ). Also, mutations that interfere with end-joining mechanisms that may be dependent on MRE11, such as the DNA-PK subunit mutant cells, are all viable (30 -33 ). These differences ultimately must reflect the usage of these genes in cell repair, recombination and broader functions. In addition, certain genes may be redundant, with other genes contributing to the same pathways such that gene mutation is not lethal.
Mre11 is complexed in mammalian and yeast cells with the product of the RAD50 gene (12 ). This association may well be functionally significant for all Mre11 roles in the cell. Because of the inability to generate mre11-/- cell lines shown here, presumably the MRE11, RAD50 and possibly XRS2 genes will produce embryonic lethality when similarly mutated by gene targeting in mouse models.
Complexes of Mre11 with Rad50 may play an essential role in some aspects of the S or G2 phases, independent of recombination functions. Alternatively, there may be a sufficiently high level of spontaneous damage that repair by Mre11-dependent homologous recombination mechanisms is always required. By analogy with functions of E.coli sbcCD, Mre11-Rad50 may associate with and be involved in the resolution of unusual DNA structures formed during DNA synthesis. sbcCD is required for resolution of cruciforms and large palindromic repeats generated during DNA replication (17 ). Purified sbcCD also has double- and single-strand exonuclease activity, a biochemical property that may be significant in its role in resolving cruciform structures. The biochemical activity of Mre11 is currently under investigation.
Human Mre11 was originally isolated in a screen for proteins that interact with DNA ligase I (23 ). Perhaps the association between these two proteins detected by two-hybrid analysis is indicative of a DNA replication role for Mre11-containing complexes. It will be interesting to learn where Mre11-deficient cells arrest in the cell cycle, as an indication of where Mre11 functions are most required.
This research was supported by an NRSA postdoctoral fellowship to Y.X. and grants from the NIH, March of Dimes Birth Defects Foundation and Novartis Pharmaceuticals to D.T.W.
REFERENCES