Title: Animals of the rattus genus wherein mshθ expression is functionally decreased and methods for the generation of specific mutants there from.

The invention relates to the field of genetics. More specifically, the invention relates to the use of a mutant animal in a method for performing a genomic mutagenesis screen.

The invention further relates to a novel mutant animal, and to the use of an animal resulting from said mutagenesis screen as an animal model for a human deficiency.

The mismatch repair system (MMR) recognizes mismatches that are incorporated during DNA replication and surpass the proofreading activity of DNA polymerase. The MMR machinery is highly conserved in evolution and the prototypic system was first elucidated in Escherichia coli (Lahue, et al. (1989) Science 245, 160-4). In this model, a mismatch is recognized by the MutS homodimer followed by binding of the MutL homodimer and activation of the repair pathway. This leads to the excision and re-synthesis of the error containing DNA strand. In eukaryotes MutS function is separated in the recognition of single-base mismatches and small insertions or deletions loops (IDL's) of one or two extra helical nucleotides by the MutSa subunit and the recognition of large IDL's of two to four nucleotides by MutSb subunit. Both subunits are heterodimers whereby MutSa consists of MSH2 and MSHθ and the MutSb of MSH2 and MSH3 (Jiricny, J. (2006) Nat Rev MoI Cell Biol 7, 335-46). Besides maintaining the genomic integrity during replication the MMR systems also recognizes different forms of DNA damage caused by genotoxic agents (Duckett et al. (1996) Proc Natl Acad Sci U S A 93, 6443- 7). Cell lines deficient for MMR are more resistant to cell death caused by alkylating agents (Claij and Te Riele (2002) Oncogene 21, 2873-9; Claij et al. (2003) Cancer Res 63, 2062-6.), antimetabolites and intrastrand

crosslinking agents (Yang et al. (2004) Cancer Cell 6, 139-50). Two hypotheses have been proposed to explain MMR mediated apoptosis. The first states that the repeated activation of MMR due to wrong nucleotide incorporation by polymerase at the site of DNA damage causes replication fork arrests, which ultimately would lead to lethality (Karran and Bignami (1992) Nucleic Acids Res 20, 2933-40).

The second hypothesis suggests that the MMR systems can function as a molecular sensor, which can directly signal the apoptotic machinery in cause of high levels of DNA damage (Fishel (1999) Nat Med 5, 1239-41).

In humans mutations in the DNA mismatch repair genes have been linked to the cause of hereditary non-polyposis colorectal cancer (HNPCC) also referred to as the Lynch syndrome, which is characterized by early-onset colon cancer (Lynch and Smyrk (1996) Cancer 78, 1149-67). Mutations in the msh6 gene have been associated with an atypical HNPCC phenotype characterized by a late onset and a high occurrence of extra colonic tumors, especially in the endometrium (Wijnen et al. (1999) Nat Genet 23, 142-4). A hallmark of MMR deficiency in HNPCC tumors is a significantly higher mutation rate, termed replication error phenotype (RER+) reflected by an elevated spontaneous mutation frequency and the instability of simple repeat length during mitosis (Ionov et al. (1993) Nature 363, 558-61). The latter phenomenon is referred to as microsatellite instability (MSI) and has proven to be a powerful tool to diagnose HNPCC tumors (Boland et al. (1998) Cancer Res 58, 5248-57).

Most of our knowledge about the MutS genes in mammals results from mouse knockout models. Complete loss of MMR recognition in the msh2-/- mouse results in a strong reduction in the survival of these mice with a median survival time of 5 — 6 months (Reitmair et al. (1996) Cancer Res 56, 3842-9). At an early stage in life these mice develop

lymphomas followed by a later onset of adenomacarcinomas in the gastrointestinal tract (Reitmair et al. (1996) Cancer Res 56, 3842-9; de Wind et al. (1995) Cell 82, 321-30). Knocking out the Msh6 gene in mice also results in a decreased life span, however, with a median survival time of 6 to 10 months this is less severe than for msh2-/-. The msh6-/- mouse predominantly develops lymphomas and tumors in the small intestine and several tumors in other tissues, like the liver and skin (Edelmann et al. (1997) Cell 91, 467-77). Deletion of the Msh3 gene in mice did not induce a cancer phenotype, however, deletion of this gene in a Msh6 deficient background led to an indistinguishable cancer phenotype from that of the msh2-/- mouse, suggesting redundancy between Msh6 and Msh3 (de Wind et al. (1999) Nat Genet 23, 359-62). Although mouse knockout models have revealed much of mammalian MMR, one drawback in modeling human HNPCC has been the differences in tumor spectra. The human HNPCC phenotype primarily shows colorectal tumors, whereas the mouse models develop tumors in the small intestine (Reitmair et al. (1996) Cancer Res 56, 3842-9; Edelmann et al. (1997) Cell 91, 467-77). Furthermore, the high occurrence of endometrial cancers found in human atypical HNPCC, especially in the case of MSH6 germ line mutations (Wijnen et al. (1999) Nat Genet 23, 142-4), is not reflected by the mouse model, urging the need for an independent model system.

Over the last decades, the rat has matured to an important genetic model system (Lazar et al. (2005) Genome Res 15, 1717-28; Smits and Cuppen (2006) Trends Genet 22, 232-40). The complete genome sequence of the rat is now available (Gibbs et al. (2004) Nature 428, 493- 521) and recently developed techniques make it possible to generate knockout rats using a target-selected N-ethyl-N-nitrosourea (ENU)-driven mutagenesis approach (Smits et al. (2006) Pharmacogenet Genomics 16, 159-69; Zan et al. (2003) Nat Biotechnol 21, 645-51).

The present methods for generating mutant rats are cumbersome. Large amounts of animals need to be screened after the introduction of alterations in the genome of a tester animal. This is mainly due to the frequency of said alterations, which are dependent on the mutagenicity of the mutagens used to introduce said alterations. An optimal alteration frequency using such animals is found to result in about 1 alteration per IxIO ⁶ or l,5xlθ ⁶ base pairs. A higher amount of mutagen results in reduced fertility and ultimately enhanced lethality of the tester animal, thereby increasing the work load that is needed for isolating an alteration in a genomic region of interest.

For example, 3,000 Fl animals, resulting from crosses between mutagenized males and untreated female mice, had to be analyzed to identify an alteration that disrupts gene function in six out of about 100 specific genes of interest, while the animals had to be kept alive until the results of the analyses were known, to allow further breeding of an animal comprising said alteration. The total amount of animals that have to be screened to identify a deleterious mutation depends on the size of the gene of interest. About 60.000 Fl animals have to be screened to identify a truncating alteration due to the introduction of a stop codon in a gene of about 1.5 kilo base pairs (kb) with a probability of 95%. With the same probability, about 10.000 Fl animals are required for identification of a truncating alteration in a gene of about 7.5 kb.

The present invention provides means and methods for significantly improving the generation of genetic rat models. It was found by the present inventors that a rat comprising an inactivating mutation in a Msh6 gene (Msh6-/-) strongly enhances the alteration frequency in genomic mutagenesis screens from about 1 alteration in every 1 to 1.5 megabasepairs (mbp) to about 1 alteration in every 500 kilopasepairs

(kbp), which is about 2 to 3 times more efficient. This enhanced alteration frequency was observed in homozygous Msh6-/- rats.

Furthermore present inventors have determined that the mutation spectrum of ENU changes in Msh6 deficient background compared to Msh6 proficient background. The percentage of G:C to A:T transitions was significantly increased in Msh6 deficient background as compared to proficient background, whereas the percentage of A:T to G:C transitions was significantly decreased. No significant change was observed for the percentages of transversions. When considering the codon usage of a representative set of rat genes, it was found that the chance of introducing a premature stop is increased by 16% (from 4.3% to 5.0%) in Msh6 deficient background. Heterozygous rats most probably would not show an increased alternation frequency, because alternations can still be recognized and repaired by the Msh2/Msh6 subunit. In vitro studies with heterozygous Msh6 mouse cell line extracts showed no significant reduction in mismatch repair efficiency (Edelmann et al. (1997) Cell 91, 467-77). In addition it was found that the homozygous mutant animals have an average life expectancy that is compatible with the mutagenesis and breeding procedure also upon administration of the mutation inducing amount of mutagen. It was found that these animals are particularly suited for the generation of genetic models.

One alteration in every 500 kilopasepairs (kbp) indicates that a rat with a genome size of 3,000,000 kbp harbors about 6.000 genomic alterations. Because the coding region (the transcriptome) of the rat encompasses about 35,000 kbp, about 70 of these alterations will reside in coding regions. On average, about 24% of these alterations are silent, about 70% result in an amino acid changes and about 6% result in the introduction of a premature stopcodon. As a consequence, on average about 50 amino acid changing alterations and 4 premature stopcodons are present in said mutagenized Msh6-/- rat. In contrast, a mutagenized wild type rat has 20 amino acid changing mutations and 1.6 premature

stopcodons, while a non-treated wild type rat has on average 0.25 alterations per animal and 0.02 stopcodons per animal.

The invention therefore provides a method for producing an animal of the genus rattus with a mutagen induced mutation in a germ line cell characterized in that said mutagen is provided to an animal of the genus rattus having a functionally reduced mshθ protein expression in said germ line cell, the method comprising providing an animal at the genus rattus having a functionally reduced mshθ expression in a germ line cell with a mutation inducing amount of a mutagen.

The present invention thus provides a method for generating an animal of the genus rattus having a mutation in a germ line cell comprising providing said animal with a mutation inducing amount of a mutagen, said method characterised in that mshθ expression in said germ line cell is functionally decreased. Said method for producing an animal of the genus rattus with a mutagen induced mutation in a germ line cell comprises providing said animal having said germ line cell with a mutation inducing amount of a mutagen, said method characterised in that mshθ expression in said germ line cell is functionally decreased. Said method thus comprises providing an animal of the genus rattus with a mutation inducing amount of a mutagen, said method characterised in that mshβ expression in said germ line cell is functionally decreased prior to or during contacting said animal with said mutagen. In other words, said method comprises providing an animal of the genus rattus, in which mshθ expression in a germ line cell is functionally decreased prior to or during contacting said animal with said mutagen, with a mutation inducing amount of a mutagen. The mutation-inducing amount of said mutagen is provided to generate a mutation in a gene other than a gene encoding Mshθ.

The invention also provides a method for providing a germ line cell of an animal of the genus rattus with a mutation comprising providing an animal of said genus comprising a germ line cell with a mutation inducing amount of a mutagen, characterised in that mshθ expression in said germ line cell is functionally decreased prior to or during contacting said animal with said mutagen.

Further provided is a method for providing a germ line cell of an animal of the genus rattus with a mutation comprising contacting a germ line cell of said animal in vitro with a mutation inducing amount of a mutagen and transplanting said germ line cell into said animal or a compatible member of the same species, said method characterised in that mshθ expression in said germ line cell is functionally decreased prior to or during contacting said germ line cell with said mutagen.

In a preferred embodiment a rattus of the species norvegicus is used and preferably a laboratory strain thereof. The term rat is preferably used herein to refer to the rattus species norvegicus.

Many different mutagens are currently known. Finding a mutation inducing amount is straightforward for the person skilled in the art. Preferably the optimal mutation inducing amount is determined in a dosage-response experiment. This amount is not necessarily the same in the treatment of whole animals or the in vitro treatment of germ line cells or precursors of germ line cells.

The increased level of mutagen induced mutations in such a germ line cell increases the chance of finding a mutation in a gene other than Mshθ. Expression of mshθ can be decreased in various ways. Mutations are preferably present in the coding region, the promoter or other transcription/translation sequences of said gene other than Mshβ.

The invention further provides a method according to the invention, further comprising generating offspring with said germ line cell, a matured derivative thereof or a nucleus thereof. The invention also provides a method according to the invention, further comprising collecting germ line cells from said animal that has been contacted with said mutagen. Preferably, a method of the invention further comprises generating offspring with said collected germ line cells. A further preferred method comprises collecting and storing cells of said offspring and/or nucleic acid of said cells, preferably comprising a germ line cell or derivative thereof.

Functionally decreased mshβ expression can be accomplished by different means. For example, RNA coding for mshθ can be provided with antisense nucleic acid to target said RNA for destruction and/or to inhibit translation of the RNA. Other methods include the induction of exon skipping to generate non-sense coding regions. Yet other methods include the insertion of transposons or retroviral vectors that disrupt the coding region or insertions/deletion that inactivate transcription through methylation. Essential in any method for functionally decreasing mshθ expression is that the level of mshθ protein in the cell is reduced.

Preferably to levels that are less than 40% of the level in the homozygous Mshβ ⁺ rat germ line cell. Preferably, the level is even lower. Preferably less than 10%, and more preferably less than 1%. In a preferred embodiment no mshθ protein is detectable in the germ line cell. The same levels of mshβ protein are preferred in case essentially all cells of said animal have a functionally decreased mshθ expression. This is preferably achieved through a genetic modification that prevents any detectable production of functional mshβ protein. Such a genetic modification is termed an inactivating mutation. Thus preferably said germ line cell and/or said animal wherein essentially all cells comprise said functionally decreased mshβ expression comprises an inactivating mutation of Mshβ. Said genetic modification or inactivating mutation preferably comprises a

modification that disrupts the coding region or inactivates the promoter of Msh6. In a particularly preferred embodiment said mutation comprises the introduction of a premature stopcodon in the coding region of Msh6. Preferably said stopcodon is introduced in exon 4 of Msh6 as described in Smits et al (2006). A method of the invention is preferably used in a genomic mutagenesis screen.

A germ line cell is a gamete or a cell that contributes to the formation of gametes. Germ line cells are functionally defined as cells that can fix mutations in the germ line of a species. In other words, that can produce an entire animal either directly or indirectly via precursors, chimeras, or fusion of gamete (nuclei). Such cells provide heritable genetic change in the germ line of an animal species. Such cells in include but are not limited to spermatogonial stem cells. In a preferred embodiment said germ line cell is a precursor of male gametes. In line with this preference it is preferred that said animal of the genus rattus is preferably a male animal.

The term "genomic mutagenesis screen" indicates the screening of a collection of mutagenized animals or progeny thereof for the presence of specific gene alterations. A genomic mutagenesis screen comprises genome-wide screens, directed at identifying alterations in all genes, and region- specific or gene-specific screens, directed at alterations in a specific chromosomal region or gene. Said screening preferably comprises the detection of a specific sequence motif, preferably comprising a stopcodon. A preferred method further comprises breeding offspring animals to segregate a mutagen induced mutation from a genetic modification that causes said functionally decreased mshθ expression.

Said screen can be saturated, indicating that the probability that said collection comprises an alteration in a particular gene is at least 95%, or semi- saturated or non-saturated. Said mutagenesis screen may or

may not be suited for targeting the DNA of specific cell organelles, like mitochondria, because it is unclear whether MMR plays a role in the DNA repair in these organelles. In vitro studies using rat liver cell extracts has shown that mismatch repair activity is present in mitochondria, however no presence of msh2 was detected at protein level (Mason et al. (2003)

Nucleic Acids Res. 31, 1052-58). This indicates that MMR in this organelle is mediated independent of the msh2/msh6 dimer, because mshθ and msh3 are unstable and rapidly degraded when the binding protein msh2 is absent (Drummond et al. (1997) Proc. Natl Acad. Sci. USA 94, 10144-49).

In a preferred aspect of the invention, said further alteration, which is additional to the alteration in the Mshθ gene, is present in the germline of said rat. For the identification of at least one further genomic alteration, mutagenized rats, preferably male rats, are preferably crossed to non-mutagenized rats, preferably female rats, to generate Fl animals. Following routine breeding schemes, these Fl rats are further crossed to generate progeny rats that are homozygous for one of said at least one further alteration. It is furthermore preferred to mutagenize adult male rats at an age of at least 8 weeks, preferably at an age of about 12-14 weeks, and to wait a full round of spermatogenesis, about 60-70 days, before crossing the rats to non-mutagenized female rats. This will result in the occurrence of further alterations in a large proportion of the resulting Fl generation.

The enhanced alteration frequency of Mshθ-/- rats results in a requirement of a reduced number of Fl animals per genome wide saturated screen. In the above described example 20.000 compared to the 50.000 wild type Fl animals. This not only reduces the number of animals and animal housing costs, but also considerably reduces total screening costs. Another advantage is the altered spectrum of modifications with a significantly higher chance for the introduction of premature stopcodons that can be obtained using a method of the invention. A further advantage

of the use of a Msh6-/- rat compared to other Msh6-/- animals in a method of the invention is the late onset of spontaneous tumorigenesis in Msh6-/- rats, which starts at about 9 months, compared to at about 3 months in Msh6-/- mice.

One of the consequences of MMR deficiency is the development of spontaneous tumors, which result in decreased survival of animals. For Msh6 knockout mice, the median survival is 6 to 10 months and for Msh2 even shorter. Animals deficient of Msh3 do not show any phenotype, probably due to redundancy between Msh6 and 3. The Msh3 and Msh6 double knockout display a comparable phenotype as the Msh2 deficient animals. A unique characteristic of the rat Msh6 knockout rat appears its markedly increased survival (median survival 14 months) as compared to the knockout mouse.

Two different mshθ-deficient mice strains have been generated, both of which have targeted exon 4 of the gene (comparable to the premature stopcodon in example 1), but on different genetic backgrounds. MshθtmlRak was generated in hybrids of C57B1/6 and WW6 (Edelmann et al., (1997) Cell 91, 467-77); MshθtmlHtr was generated in hybrids of 129/OLA and FVB (de Wind et al., (1999) Nat. Genet. 23, 359-63). These models were found to display clear phenotypic differences. For example, the median survival of MshθtmlRak mice was found to be 10 months (Edelmann et al., (1997) Cell 91, 467-77), whereas that of MshθtmlHtr mice was only 6 months (de Wind et al., (1999) Nat. Genet. 23, 359-63).

The reason why the mshθ-/- rat shows a delayed onset of tumorigenesis (median survival of 14 months) when compared to the Msh6-/- mouse models remains unclear. The median life span of wild type mice and rats do not differ significantly, suggesting that other species- specific characteristics could underlie the observed differences. As rats and mice have diverged for about 40 million years this is not unlikely.

Alternatively, because of its size, the rat could be able to sustain the growth of malignancies longer, allowing tumors to reach relatively larger sizes before the animal becomes moribund. The predominant development of lymphomas in the mshθ-deficient rats is also observed in mouse strains lacking mshθ or other MMR components (de Wind et al., (1995) Cell 82, 321-30; de Wind et al., (1999) Nat. Genet. 23, 359-63; Edelmann et al., (1997) Cell 91, 467-77; Reitmair et al., (1995) Nat. Genet. 11, 64-70). The extensive involvement of the small intestine in the neoplastic phenotype of the MshδtmlRak mice (Edelmann et al., (1997) Cell 91, 467-77) was not observed in the rat and could potentially explain the shorter life span of mah6- deficient mice. However, the MshβtmlHtr strain rarely develops intestinal tumors, but does show the highest reduction in life span (de Wind et al., (1999) Nat. Genet. 23, 359-63). Interestingly, the Msh6-/- rat shows a high incidence of endometrial cancers (3 of 7 Msh6-/- females), resembling the atypical HNPCC spectrum of tumors. Although cancer in the uterus in mice has been reported in the MshβtmlHtr strain (3 in 22 Msh67- mice), the other msh6-deficient strain completely lacked involvement of the uterus in their cancer phenotype (de Wind et al., (1999) Nat. Genet. 23, 359-63; Edelmann et al., (1997) Cell 91, 467-77), indicating a considerably higher occurrence in the rat.

Treatment of MMR deficient animals with carcinogens that damage the DNA, such as alkylating drugs like ENU, result in hypersensitivity and a dramatic decrease of survival. After ENU treatment of Msh6 knockout rats, survival is clearly decreased (median survival 8.5 months), but in contrast to the mouse, this prolonged survival provides sufficient time to generate offspring from ENU-treated Msh6 knockout males (animals are mutagenized at 3 to 3.5 months of age and require a 2.5 month time window to go through a full round of spermatogenesis and generate genetically stable offspring). Hence, more than half of the mutagenized animals can be used for generating mutant

Fl offspring for on average 2 months, which would not be possible in any of the MMR-deficient mice models.

Furthermore, a Msh6-/- zebrafish (Danio rerio) was found not to result in an enhanced alteration frequency following exposure to a mutagen (Feitsma et al., (2008) Mutagenesis 23, 325-29). This could be explained by the large difference in mutation frequency. In wild type zebrafish, the ENU-induced mutation frequency is about 1 mutation every 100,000 bp, whereas, this frequency in wild type rats is more than 10-fold lower (Smits et al., Pharmacogenet. Genomics (2006) 16, 159-69). It is possible that the zebrafish mutation load is the maximum that is compatible with viability, a suggestion that is corroborated by comparable maximal mutation frequencies observed in C. elegans (Cuppen et al., (2007) Genome Res. 17, 649-58) and Arabidopsis (Till et al., (2003) Genome Res. 13, 524-30). The lower ENU-induced mutation frequencies in rodents could at least partially be explained by a more efficient mismatch repair in the testis that counteracts mutagenicity, and is less likely due to increased sensitivity to genotoxic damage in general. The present invention for generating progeny using cells wherein mshδ-expression is functionally decreased is also suitable for the generation of plants cells having mutation as a result of exposure to a mutation inducing amount of a mutagen.

An inactivating mutation in a Msh6 gene preferably results in a functionally disabled mshδ protein. Said mutation is any genomic or epigenomic alteration that results in the production of an abnormal, nonfunctional msh6 protein, a partially active version of the protein, a reduced level of expression of said protein, or the absence of mshδ protein. A functionally disabled mshδ results in an increased number of genomic mismatches that are left unrepaired during cell division. The mismatch repair protein mshδ, in a heterodimer with msh2, forms the mammalian MutS complex, which is part of the mismatch repair (MMR) system. The

MMR system is responsible for removal of base mismatches caused by spontaneous and induced base deamination, oxidation, methylation and replication errors. The MutS complex recognizes base mismatches and forms an active complex with MLH1-PMS2 proteins for actual repair of the mismatch after recognition. Methods to determine a deficiency in the MMR system are known to a skilled person and include, but are not limited to, methods for determining microsatellite instability (MSI) in simple repeats, and methods for determining the frequency of point mutations.

In a preferred embodiment, said inactivating mutation results in a premature stop codon in the coding part of said Msh6 gene. Mutations in the human MSH6 gene have been reported in about 10 percent of hereditary nonpolyposis colon cancer (HNPCC) families. A high percentage of these mutations result in premature stop codons, causing the production of an abnormally short, nonfunctional MSH6 protein. Therefore, the presence of a premature stop codon closely resembles the frequently occurring functionally disabled human MSH6 protein.

In a most preferred embodiment, said inactivating mutation comprises the mutation of a codon for a leucine at amino acid position 306 of the protein to a stop codon. Said mutation resides between the PCNA binding domain and the mismatch-binding domain and results in nonsense-mediated decay of the mRNA transcript. The mshθ protein consists of a (1) PCNA domain that is needed for connecting the protein to DNA replication, a (2) mismatch-binding domain, a (3) connector domain that is involved in allosteric signaling between the levers domain and the ATPase domain, the (4) levers domain, which is most probably involved in signal transduction between the ATPase and DNA binding domains, the (5) clamp domain that make significant nonspecific DNA contact and the (6) ATPase domain (Warren et al. (2007)Mol Cell 26, 579-92). Alterations that influence the structure of any of these characterized domains will

most probably influence the protein function in one way or another. Premature stops in most cases will lead to nonsense-mediated decay of the mRNA and will completely abolish protein function.

A further alteration can be introduced in Msh6-/- rats by any means known to a skilled person, comprising radiation such as ionizing radiation and ultraviolet light, insertional mutagenesis, and chemical mutagenesis. A mutagen induced mutation is preferably a point mutation. Preferably a transversion or a transition.

A preferred mutagen in a method of the invention comprises a chemical mutagen.

A chemical mutagen in general causes alterations in the DNA of cells that are repaired by the MMR system. A deficiency in the MMR system will result in an enhanced alteration frequency following exposure to a chemical mutagen. Known chemical mutagens comprise a base analogue such as bromouracil and aminopurine,; an intercalating agent such as acridine orange, proflavin, and ethidium bromide; a deaminating agent such as nitrous acid; and an alkylating agent, such as a methylating or ethylating agent.

A preferred chemical mutagen comprises an alkylating agent, such as, for example, nitrosoguanidine, methyl methanesulfonate, ethyl methanesulfonate (EMS) and, most preferred, N-ethyl-N-nitrosourea (ENU) or N-methyl-N-nitrosouera (MNU).

ENU is an alkylating agent that is a powerful mutagen in spermatogonial stem cells. Therefore, the use of ENU as a chemical mutagen results in a high percentage of further alterations that are present in the germline of said mutagenized rat.

ENU predominantly modifies A:T base pairs, with 44% A:T- >T:A transversions, 38% A:T->G:C transitions, 8% G:C ->A:T transitions, 3% G:C->C:G transversions, 5% A:T->C:G transversions and 2% G:C->T:A transversions, resulting in about 64% missense mutations, 10% nonsense mutations and 26% splicing errors (Justice et al. (1999) Hum MoI Gen 8: 1955-1963). Therefore, the use of ENU as a mutagen will result in about 4 to 5% of said further alterations introducing a premature stop codon as a result of a nonsense mutation.

Said chemical mutagen can be administered orally, intravenously, intraperitoneally, or directly injected into the testis. However, care is to be taken that the solvent used does not interfere with the route of administration. Furthermore, some routes of administration are not applicable for all mutagens, as is known to a skilled person. For example, ENU is not effective if orally applied to rats and is preferably administered intrapreritonally.

The use of a mutagen in an animal of the genus rattus, comprising cells wherein mshδ expression is functionally decreased, results in mutations that are different in different cells of the animal. The average mutation frequency in male germ line cells as calculated in the experimental section is markedly elevated when compared to a wild type animal. Thus the invention further provides an animal of the genus rattus comprising a cell with a mutagen induced mutation characterised in that msh6 expression in said cell is functionally decreased in addition to said mutagen induced mutation. Preferably said animal of the genus rattus comprises a plurality of cells wherein mshδ expression is decreased, preferably said plurality of cells comprise different mutagen induced mutations. Preferably essentially all cells of said animal comprise a functionally decreased mshθ expression. With essentially is meant that some cells also normally do not express mshθ. For instance mature red blood cells that do no longer contain a nucleus and skin cell that stop RNA

synthesis as they mature and move upward in the skin. The observed mutation frequency is higher than any mutation frequency previously observed in animals that are capable of generating offspring. An animal of the invention thus preferably comprises an inactivating mutation of mshθ in essentially all cells and at least one mutagen induced mutation in every 0.5 x 10 ⁶ to 1 x 10 ⁶ base pairs. This animal can be the mutagen treated animal, or offspring thereof. Offspring is typically generated with a normal counterpart, i.e. not exposed to a mutagen. In this offspring one set of chromosomes, e.g. derived from the mutagen exposed animal comprises the mutagen induced mutations with at least the indicated frequency while the other set of chromosomes does not. In a preferred embodiment the same chromosome in each cell comprises the same (frequency) of mutagen induced mutation. Thus the present invention further provides an animal with a germ line cell that comprises at least one mutagen induced mutation in every 5 x 10 ⁶ base pairs on at least one set of chromosomes. Preferably said germ lines cells of said animal comprise on average at least one mutagen induced mutation in every 5 x 10 ⁶ base pairs. Preferably essentially all cells of said animal comprise a functionally decreased mshθ expression. An animal of the invention preferably comprises a plurality of germ line cells comprising the mentioned mutagen induced mutation, wherein at least two and preferably at least 5 germ line cells comprise different mutagen induced mutations.

Using a method of the invention, many different mutations were generated and detected in progeny rats. Many of these mutations are listed in table 7, table 8 and table 11. In a preferred embodiment an animal of the invention comprises a mutation in a gene as listed in table 7, table 8 or table 11. The mutant of a gene listed in table 7 and 11 preferably comprises the mutation as specified in table 7 and 11. The animals can be used directly to generate a genetic model. However, typically the rat having an

interesting mutation is crossed into a suitable background that typically no longer comprises the genetic modification resulting in decreased mshθ expression, although this does not necessarily need to be the case. Double or further mutants can be of interest also, for instance as a model for HNPCC.

Thus the present invention further provides progeny of an animal according to the invention, having a mutation in a gene as listed in table 7, 8 or 11 and comprising cells having functional mshθ expression. Crossing may result in the generation of animals that are homozygous for a mutation in a gene of table 7, 8 or 11. Such homozygotes typically, though not necessarily have the same mutation in each of the copies in the cell. Thus preferably the invention provides an animal of the invention that is homozygous for said mutation in a gene listed in table 7, 8 or 11, whereby said animal does not comprise a cell wherein mshθ expression is functionally decreased. Said mutation in said animal is preferably characterized by a G:C to A:T transition, as this transition was significantly increased in mshβ deficient background as compared to a proficient background, while an A: T to G: C transition was significantly decreased in MSHθ deficient background. No significant change was observed for the percentages of transversions.

The invention further provides an animal of the genus rattus comprising a cell with a mutagen-induced mutation which animal is a progeny of an animal according to the invention. Preferably said animal comprises a cell wherein mshβ expression is functionally decreased, in addition to said mutagen-induced mutation. The invention further provides an animal of the genus rattus comprising a cell with a mutagen induced mutation and wherein essentially all cells comprise a functionally decreased mshθ expression, in addition to said mutagen-induced mutation.

The identification of said at least one further alteration in progeny of a mutagenized animal can be performed according to any

method known in the art. Phenotypic screens can be performed for the identification of phenotypic alterations in progeny of said mutagenized animal. The presence of a dominant alteration that results in a visible phenotype may be scored already in the Fl generation, while a breeding scheme allows scoring for the presence of a recessive alteration including a developmental lethal alteration. A phenotype that, for example, affects the gross anatomy, the coat, or the size of an animal can be easily identified. Furthermore, X-ray analysis, blood tests, and/or specific tests such as behavioral tests, learning test and/or memory test can be introduced in the screening program to identify alterations that affect the outcome of these tests. In addition, microarray analyses can be used to identify changes in gene expression patterns in a specific tissue of during a specific developmental stage.

Therefore, a preferred method according to the invention further comprises breeding said offspring animals to segregate said mutagen induced mutation from a genetic modification that causes said functionally decreased Msh6 expression.

A breeding scheme can also be applied to identify modifiers of a known pathway by introducing a rat model of a human disease and identifying genomic alterations that result in an amelioration or cure of said model disease. Alternative, a marker gene can be introduced in the breeding scheme and the progeny rats can be screened for alterations that modify, for example, the expression or localization of said marker gene.

Further to the phenotypic screens, or in addition thereto, gene- or region- specific screens can be performed to isolate alterations in a specific gene or genes, or in one or more genomic regions. A preferred method according to the invention further comprises screening offspring of said animal for the presence of a mutagen induced mutation in a specific nucleic acid sequence. Following the isolation of genomic DNA from a

progeny of the mutagenized animals, the genomic regions of interest can be amplified, for example by cloning, polymerase chain reaction (PCR) and/or multiplex PCR, and an alteration in the amplified material, as compared to the progeny of a non-mutagenized rat, can be determined. Methods for high-throughput single nucleotide polymorphism detection are known in the art, including DNA resequencing techniques, conformation- sensitive capillary electrophoresis, high-resolution "melt" analysis, sequence analysis such as array-based sequence analysis. Targeting induced local lesions in genomes (TILLING) approaches can be used for combining chemical mutagenesis with a sensitive DNA screening- technique in a target gene such as, for example, HPLC or the restriction enzyme CeI-I combined with a gel based system to identify mutations.

The invention also provides the use of a rat comprising an inactivating mutation in a Msh6 gene, in a genomic mutagenesis screen. The use of said rat will result in an enhanced alteration frequency, enabling a reduced number of animals for said screen. Said mutation preferably results in a functionally disabled Msh6 protein such as the introduction of a premature stop codon in the coding part of said Msh6 gene. A preferred mutation is a mutation of a codon for a leucine at amino acid position 306 of the protein to a stop codon.

In a further embodiment, the invention provides a male rat, or progeny there from, comprising an inactivating mutation in a gene encoding Msh6, said rat being exposed to an effective dose of a mutagen. It is preferred that said rat comprises at least one further alteration in the germ line as compared to a litter mate that was not exposed to said mutagen. Methods for determining said at least one further alteration are known in the art, including DNA resequencing, conformation- sensitive capillary electrophoresis, high-resolution "melt" analysis and sequence analysis.

The invention further provides sperm that is isolated from a rat of the invention or male progeny derived there from. Methods for isolating sperm from rats are known to a skilled person. For example, sperm can be directly isolated from the caudal epididymides or vas deferens from an anesthetized male. If required, sperm cells such as spermatozoa can be isolated from said sperm. Sperm or sperm cells can be preserved by cryopreservation applying cryoprotection methods to prevent damage during freezing and thawing. Alternatively, freeze-drying or desiccation by rapid convective drying (McGinnis et al. (2005) Biol Repro 73, 627-633) can be applied for storage of rat sperm or sperm cells.

The invention furthermore provides the use of sperm of the invention for identifying said at least one further genomic alteration in progeny derived from said sperm. Sperm according to the invention can be used to fertilize rat oocytes to establish progeny from said sperm. If required, said stored sperm can be injected into oocytes by the intracytoplasmic sperm injection technique (ICSI) to achieve fertilization.

The invention also provides the use of in vitro mutagenesis of spermatogonial stem cells (SSC) or embryonic stem (ES) cells that have reduced mshδ function, e.g. by derivation of such cells from Msh6-/- animals. Methods for isolating SSC or ES cells are known to a skilled person. SSC and ES cells can be mutagenized and cryopreserved or transplanted to recipient male testis or blastocysts, respectively. In the first case Fl progeny can be produced by breeding with a wild-type female, in the second case the injected blastocyst can be injected in recipient uterus and give rise to chimaeras, which can be bred with wild-type animals to preserve alternations.

In another aspect, the invention provides a method of selecting a rat homozygous for one of said at least one further mutation, comprising breeding a rat according to the invention or using sperm according to the

invention, and selecting a rat comprising a wildtype Msh6 gene and being homozygous for one of said at least one further mutation. Said further mutation in said animal is preferably characterized by a G:C to A:T transition, as this transition was significantly increased in MSH6 deficient background as compared to a proficient background, while an A: T to G: C transition was significantly decreased in mshβ deficient background. No significant change was observed for the percentages of transversions.

A rat according to the invention comprises an inactivating mutation in a Mshθ gene and at least one further alteration. It is possible that the phenotype of one of said further alterations, preferably mutations, either heterozygous or homozygous, is affected by said mutation in a Mshθ gene and/or one of said further alterations. Said method of the invention allows the generation of a heterozygous or homozygous rat that is mutant for only one of said further alterations, in the absence of an inactivating mutation in a Mshβ gene.

In a preferred embodiment said further alteration is an alteration in a gene as listed in table 7, 8 and 11. The alteration or mutation is preferably an inactivating mutation. Preferably a mutation in the coding region of said gene of table 7, 8 and 11. Preferably said alteration is an alteration as listed in table 7 and 11.

The invention in addition provides the use of a rat comprising an inactivating mutation in a gene encoding Mshθ as an animal model for HNPCC. Said rat is characterized by microsatellite instability in simple repeats and point mutation instability, has a reduced life span, and develops tumors in the HNPCC spectrum. In contrast to established mouse Mshθ knockout models, said Mshθ-/- rat lacks intestinal tumors, thereby more closely resembling the spectrum of tumors seen in human HNPCC patients. The Mshβ-/- rat does not develop colorectal tumors but

develops an atypical HNPCC spectrum of tumors, especially characterized by the presence of endometrial cancers.

The invention in addition provides the use of a rat according to the invention for testing a potential pharmaceutical therapeutic agent. The extended range of phenotypic characteristics, combined with the size of the animal, the prolonged viability and the capacity to bear big sized tumors make this rat an attractable model for specific experimental manipulations, such as for example local irradiation experiments for studying and treating residual cancer cells.

The invention further provides a method for providing a rat cell with a mutation comprising exposing said rat cell to a mutation-inducing amount of a mutagen, said method characterized in that Msh6 expression in said rat cell is functionally decreased prior to or during exposing said cell to said mutation-inducing amount. Preferably said rat cell is a germ line cell or derivative thereof. As mentioned herein above, embryonic or spermatogonial stem cells can be exposed in vitro, to a mutagen inducing amount of a mutagen, and be used to generate a rat, for instance via insertion into a treated blastocyst or testis. The ES cells can also be characterised for the presence of the appropriate mutation before generating said animal. The ES cell can be generated in the traditional way or be generated by dedifferentiation from a differentiated cell. For instance, it is possible to generate ES like cells from adult cells through the expression therein of three or four genes. Another method is the use of a spermatogonial stem cell. A derivative of a germ line cell is preferably a cell that is the result of transfer of a nucleus of said rat cell into a rat germ line cell. The invention further provides a rat cell comprising a mutagen induced mutation that is obtainable by a method of the invention. Said rat cell preferably further comprises a genetic modification resulting in a functionally decreased expression of mshθ. Preferably said rat cell is a germ line cell or a derivative thereof.

The invention further provides a collection of rat cells characterized in that said collection comprises at least one cell obtainable by a method according to the invention. Said rat cells are preferably present in a plurality of receptacles. Said cells can be used a source for screening for a desired mutation. Keeping cell in storage is more practical than maintaining unidentified mutations in life rats. Thus preferably said collection is used for identifying a mutation that is induced by said mutagen. In a preferred embodiment said collection is stored in a storage means. Preferably a storage means for long term storage such as storage in frozen form at reduced temperatures such as -70 degrees Celsius. The invention further provides the use of a rat cell or a nucleic acid containing derivative thereof from a collection of the invention for the generation of a rat.

In a further aspect the invention provides a collection of germ lines cells or derivatives thereof comprising a mutagen induced mutation and comprising an msh expression that is functionally decreased. Preferably said collection comprises a plurality of germ line cells or derivatives thereof comprising at least two and preferably at least 5 germ line cells or derivatives thereof with a different mutagen induced mutation.

Figure legends

Figure 1. Characteristics of the mshβ '- rat. (A) Genomic organization of the mshβ gene in rats. The arrow marks the position of the ENU-induced mutation that results in a premature stop codon. The sequence trace indicates the T to A transition in a heterozygous rat. (B) Western blot (WB) of testes and cultured rat embryonic fibroblast (CREF) lysates stained with antibodies against human MSH6 (160 kDa). Lysate of human embryonic kidney (HEK) cells was loaded as a positive control. MSH6 protein is completely absent in mshβ-'- testes and CREF. Coomassie brilliant blue staining (CBB) was used as a control for protein loading.

Figure 2. Microsatellite instability in the germ line. (A) A mononucleotide repeat of 20 repetitive subunits (G)2o was analyzed in a cross between a mshβ' ⁱ - father and a wild type mother and their mshβ ^+/ - offspring. The red line indicates the size of the most prominent amplification product in both parents. In the DNA sample of the offspring one of the alleles has lost one repetitive subunit. (B) Fragment analysis of a dinucleotide repeat (CA)3β that is polymorphic between the founder animals. As a result, progeny is heterozygous for this marker and the depicted Fl animal shows a deletion of one repetitive subunit in the larger allele, which was passed on by the mshβ '- father. The orange peaks represent the size marker that was loaded to determine the size of the PCR product.

Figure 3. Survival of the mshβ '- rats. The time the rats became moribund is recorded. Red line, mshβ '' rats (n=17); black line, control group (n=100, Source Harlan).

Figure 4. Tumors observed in the mshβ '' rat. (A) Section of a mediastinal lymphoblastic lymphoma of a mshβ '- male rat stained with HE. (B) A higher magnification of the lymphoblastic lymphoma. (C) Section of a leiomyosarcoma of the endometrium stained with HE. (D) Section of an

endometrial carcinoma stained with HE. (E) A higher magnification of the endometrial carcinoma. (F) A section of a squamous cell carcinoma in the stomach of a mshβ ^ female rat stained with HE. (G) Section of a fibroadenoma of a mshβ '- male rat stained with HE. (H) Section of a liver with lymphoblastic lymphoma leukemia stained with HE. (I) Section of a testicle containing a Leydig cell tumor stained with HE.

Figure 5. Survival of mshβ ^ male rats after treatment of different concentrations of ENU compared to untreated mshβ '- male rats. Time points of the last ENU injection and the beginning of the mutation discovery screen are indicated.

Figure 6. ENU-induced target- selected knockout protocol. (A) Mshβ ¹ - males were mutagenized and crossed with wild-type Wistar females. At two weeks of age tissue samples from the Fl progeny were taken and DNA was isolated. (B) For every DNA samples 768 pre-selected amplicons were PCR amplified and dideoxy sequenced. (C) Sequence traces for the same amplicon of the different animals was aligned, processed with PolyPhred (Nickerson et al, (1997) Nucleic Acids Res 25, 2745-51) and analyzed using in house developed software. (D) ENU-induced de noυo mutations were verified by independent PCR amplification and sequencing and pups carrying interesting mutations were weaned. The whole procedure from taking tissue samples to weaning interesting mutants takes 1 week, indicated in the centre of the figure.

Figure 7. ENU-induced mutation spectrum in MSH6 deficient background compared to MMR proficient animals. (A) The spread of ENU- induced transitions in MSH6 proficient and deficient background differs significantly, but the spread of ENU-induced transversions in MSH6 proficient and deficient background does not differ. (B) The change of introducing a premature stop codon with ENU-induced MSH6 proficient and deficient mutation spectra. For the calculations the sequences of the

768 representative exon-containing amplicons were used (* indicates significance, p < 0.05).

Figure 8. Probability of gene knockouts in MSH6-deficient and wild type rats

The chance to retrieve a knockout for any given gene and the total number of genes that will be knocked out when all genes would be screened for mutations is plotted as a function of the number of mutant Fl animals for wild type (WT, 40 mg ENU/kg, black line) and MMR- deficient (Msh6-/-, 30 mg ENU/kg, red line) rats. The red dashed lines show the number of animals needed to knockout 95% of all genes and any given gene with 95% chance. The use of a mshθ-deficient background reduces the number of animals ~2.5-fold.

Examples

Example 1

Materials and Methods

Animals

All experiments were approved by the Animal Care Committee of the Royal Dutch Academy of Science according to the Dutch legal ethical guidelines. Experiments were designed to minimize the number of required animals and their suffering. The Msh6 knockout rat (Msh6 ^1Hubr ) was generated by target- selected ENU-driven mutagenesis (for detailed description, see (Smits et al, (2006) Pharmacogenet Genomics 16, 159-69). Briefly, high- throughput resequencing of genomic target sequences in progeny from mutagenized rats revealed an ENU-induced premature stop codon in exon 4 of the Msh6 gene in a rat (Wistar/Crl background). The heterozygous mutant animal was outcrossed at least three times to eliminate confounding effects from background mutations induced by ENU. To obtain homozygous animals the heterozygous offspring was crossed in. At three weeks of age ear cuts were taken and used for genotyping. Genotypes were reconfirmed after experimental procedures were completed. Animals were housed under standard conditions in groups of two to three per cage per gender under controlled experimental conditions (12-h light/dark cycle, 21±1°C, 60% relative humidity, food and water ad libitum).

Western Blot analysis

Proteins were extracted by adding lysis buffer (1% SDS, 1.OmM sodium ortho-vanadate, 1OmM Tris pH 7.4) to approximately 0.25 g of tissue or cultured cells. After incubating the homogenate at 100 ⁰ C for 1 min the supernatant fluids were obtained after 5 min of centrifugation at 14,000 X g at room temperature. One volume of SDS loading buffer (125 mM TRIS- CL pH 6.8, 3% SDS, 10% glycerol, 10 mM DTT and 0.1% bromophenol

blue) to one volume of lysate and then incubated for 5 min at 100 ⁰ C. The protein was separated on a SDS gel (6% acrylamide gradient, Bio-Rad) and transferred to a nitrocellulose membrane. Non-specific, protein- binding sites were blocked by incubating the membrane with blotting buffer: 0.05% PBS-Tween containing 5% nonfat dry milk for 1 hour at room temperature. The membrane was washed with 0.05% PBS-Tween and incubated overnight at 4°C with a 1:100 dilution of a monoclonal mouse anti-human MSH6 antibody (BD Biosciences Pharmingen) in blotting buffer. The membrane was washed five times for 10 min in 0.05% PBS-Tween and then incubated for 2 hours with peroxidase-conjugated, anti-mouse IgG diluted 1:2500 in blotting buffer at room temperature. The membrane was washed five times with 0.05% PBS-Tween and protein bands were detected by using the enhanced chemiluminescence detection method (ECL, Amersham Biosciences). Cell lysate from human HEK cells was used as a control and total protein was measured by Coomassie Brilliant Blue staining.

Genotyping

Lysis on ear cuts was done overnight at 55°C in 400 μl lysis buffer, containing 100 mM Tris-HCL (pH 8.5), 200 mM NaCl, 0.2% SDS, 5 mM EDTA and 100 μg/ml of freshly added Proteinase K. Proteinase K was inactivated by incubating 10 minutes at 95 ⁰ C, followed by cooling on ice. Genomic DNA was isolated by adding one volume of isopropanol, mixing and centrifugation for 40 minutes at 3000 x g at 4°C. The supernatant was discarded and the samples were rinsed with 70% ethanol and dissolved in 500 μl H2O. Genotyping was performed using the KASPar SNP Genotyping System (KBiosciences, Hoddesdon, UK) and gene specific primers (forward common, CAG TGG ACC CAC TAT CTG GTA; reverse al, GAA GGT GAC CAA GTT CAT GCT CTT CTC TGG CTT AAG CCA TTC TA; reverse a2, GAA GGT CGG AGT CAA CGG ATT CTC TTC TCT GGC TTA AGC CAT TCT T). Shortly, a PCR was carried out using the optimal thermocycling conditions for KTaq (94°C for 15 min; 20 cycles of

94°C for 10 sec, 57°C for 5 sec, 72°C for 10 sec; followed by 18 cycles of 94°C for 10 sec, 57°C for 20 sec and 72°C for 40 sec; GeneAmp9700, Applied Biosystems, Foster City CA, USA). The PCR reaction contained 2 μl DNA solution, 1 μl 4X Reaction Mix, 165 nM reverse Primer al and a2 and 412.5 nM of the common forward primer, 0.025 μl KTaq polymerase and 0.4 mM MgC12 in a total volume of 4 μl. Samples were analyzed in a PHERAstar plate reader (BMG Labtech) and data was analyzed using Klustercaller software (KBiosciences). All genotypes were confirmed in an independent reaction.

MSI analysis

A simple repeat containing a repetitive stretch of (CA)40 (chrl4:85120966- 85121141, RGSC 3.4) was amplified using the following primers: 5'-6FAM- TTC AAC CAC AAT CTC GAC AG -3' (forward) and 5'- AGG CAT GAG TTC TGA GGT TC -3' (reverse); for the simple repeat (contig) with the stretch of (CA)36 (chrl3: 105939209- 105939433) was amplified using the following primers: 5'-6FAM- TGG CAC AGG TGT TTA GTG TC -3' (forward) and 5'- TGC AGA AGA AAT GAG AGG TG -3' (reverse); for amplifying the simple repeat containing a (G)20 (chr3: 105633672- 105633851) 5'-VIC- CAT TCT GGA AGT GAC TGC TG -3' and 5'- TCC ACG ATA CTG CAA TTC TC (reverse) were used and the simple repeat containing the stretch of (A) 30 (chr8: 118654926- 118655555) was amplified using the following primers 5'-NED- GCC CTC TTC TGG TGT ATC TG -3' (forward) and 5'- AGC TTC ATC CGT TAG TGT GG -3' (reverse). The PCR was carried out using a touchdown thermocycling program (94°C for 60 sec; 15 cycles of 92 ⁰ C for 30 sec, 65 ⁰ C for 30 sec with a decrement of 0.6°C per cycle, 72 ⁰ C for 60 sec; followed by 30 cycles of 92 ⁰ C for 30 sec, 58°C for 30 sec and 72 ⁰ C for 60 sec; 72 ⁰ C for 180 sec; GeneAmp9700, Applied Biosystems). The PCR reaction contained -100 ng DNA, 0.2 μM of each forward primer and 0.2 μM of each reverse primer, 400 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% dimethyl sulfoxide (DMSO) (w/v), 2 mM MgC12, 85 mM ammonium acetate pH 8.7 and 0.4 U Taq

Polymerase in a total volume of 10 μl. All the PCR products were tested on a 1% agarose gel containing ethidium bromide for the presence of the proper PCR fragment. 0.5 μl of GeneScanTM-500 LIZTM size standard (Applied Biosystems) was added to every PCR product and diluted to an end volume of 10 μl with H2O. The samples were analyzed on a 96- capillary 3730XL DNA analyzer (Applied Biosystems), using the fragment analysis protocol on 36 cm array. Simple repeat sizes were analyzed for the presence of heterozygous size polymorphism using GeneMapper software (Applied Biosystems). All samples were analyzed twice in independent PCR reactions and fragment analyses.

Point mutation instability analysis

768 pre-selected amplicons were amplified using a nested PCR setup. The first PCR was carried out using a touchdown thermocycling program (94°C for 60 sec; 15 cycles of 92°C for 30 sec, 65°C for 30 sec with a decrement of 0.6 ⁰ C per cycle, 72°C for 60 sec; followed by 30 cycles of 92°C for 30 sec, 58°C for 30 sec and 72°C for 60 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems). The PCR reaction contained about 100 ng DNA, 0.2 μM of each forward primer and 0.2 μM of each reverse primer, 400 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% dimethyl sulfoxide (DMSO) (w/v), 2 mM MgC12, 85 mM ammonium acetate pH 8.7 and 0.2 U Taq Polymerase in a total volume of 5 μl. After thermocycling, the PCRl reactions were diluted with 20 μl H2O and 1 μl was used as template for the second round of PCR. The second PCR was done using a standard thermocycling program (94°C for 60 sec; 35 cycles of 92°C for 20 sec, 58°C for 30 sec and 72°C for 60 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems). PCR2 mixes contained 1 μl diluted PCRl template, 0.1 μM forward primer, 0.1 μM reverse primer, 100 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% DMSO (w/v), 2 mM MgC12, 85 mM ammonium acetate pH 8.7 and 0.1 U Taq Polymerase in a total volume of 5 μl. Several samples of each plate were tested on a 1% agarose gel containing ethidium bromide for the presence of the proper PCR fragment.

PCR2 products were diluted with 20 μl H20 and 1 μl was directly used as a template for the sequencing reactions. Sequencing reactions contained 0.1 μl BigDye (v3.1; Applied Biosystems), 1.9 μl BigDye Dilution Buffer (Applied Biosystems) and 0.4 μM of the forward primers used for the PCR2 reaction in a total volume of 5 μl. The thermocycling program that was recommended by the manufacturer was used for the sequence reactions (40 cycles of 94°C for 10 sec, 50 ⁰ C for 5 sec and 60 ⁰ C for 120 sec). Sequencing products were purified by ethanol precipitation in the presence of 40 mM sodium-acetate and analyzed on a 96-capillary 3730XL DNA analyzer (Applied Biosystems), using the standard RapidSeq protocol on 36 cm array. Sequences were analyzed for the presence of heterozygous mutations using PolyPhred (26) and in-house developed software. All candidate mutations were verified in independent PCR and sequencing reactions.

Analysis of tumors

Animals were scarified by CO2/O2 suffocation. Tumors, if found, and organs including the gastrointestinal tract, lungs, liver, kidneys, spleen and thymus were removed and fixed in phosphate buffered 4% formaldehyde. Representative tissues from the tumors and organs were processed and embedded in paraffin. All tissues were prepared for Hematoxylin and Eosin stain.

Results

Generation of the Msh6 knockout rats

In a large ENU-driven target- selected mutagenesis screen we identified a rat carrying a heterozygous mutation in the Msh6 gene (Smits et al, (2006) Pharmacogenet Genomics 16, 159-69). The mutation, a conversion of a T into an A in exon 4, resulted in a premature stop codon at position 306

(Figure IA). After outcrossing the mutant rat, heterozygous offspring was used to obtain homozygous mutant rats, which occurred in a Mendelian

fashion (not shown). The homozygous mutant rats were viable and showed normal growth and fertility. Confirmation of the Msh6 gene knockout phenotype at the molecular level was established by Western blotting, showing that the full-length Msh6 protein of approximately 160 kDa in the testes of Msh6+/+ and Msh6+/- males was completely lacking in the testes of Msh67- males (Figure IB). We could not detect any truncated protein, suggesting that either the mRNA with the premature stop codon or the truncated protein are unstable, due to nonsense-mediated decay or protein folding defects, respectively. Cell lysates of cultured rat embryonic fibroblasts also showed complete absence of Msh6 in homozygous mutant cultures, further confirming these results.

Msh6 deficient rats show microsatellite instability in the germ line One of the roles of Msh6 in the MMR pathway is the recognition of small insertions or deletions loops (IDL's) of one or two extra helical nucleotides (Jiricny (2006) Nat Rev MoI Cell Biol 7, 335-46). Deletion of Msh6 is therefore expected to result in the inability to repair simple sequence length polymorphisms that occur during DNA replication and result in microsatellite instability (MSI) at mono- and dinucleotide repeats. MSI was analyzed in the out crossed offspring of 9 Msh6-/- males and 19

Msh6+/+ males by assaying two mononucleotide repeats ((G)20 and (A)30) and two dinucleotide repeats ((CA)36 and (CA)40). In both mononucleotide repeats (Figure 2A) and dinucleotide repeats (Figure 2B) MSI was observed as a mono-allelic change of the length of the tested repeat. Although for some of the repeats the paternal (-/-) and maternal (+/+) contribution could not be distinguished (Figure 2A), in other cases the size of the repeat was polymorphic within the strain and showed that the MSI phenotype is exclusively contributed by the paternal allele (Figure 2B). In total 168 progeny of the 9 homozygous mutant males were tested and revealed that germ line MSI occurred in all males and in all the repeats tested (6%, ± 4% for (CA)36; 9%, ± 2% for (CA)40; 8%, ± 2% for (G)20; 10%,

± 7% for (A)30; see table 1). None of the 100 offspring samples from 19 control wild-type males and females showed MSI.

Msh6 knockout rats show a germ line mutator phenotype Besides recognizing length polymorphisms in simple repeats the Msh6 protein is also involved in the recognition of single-base mismatches introduced during DNA replication. The inability to perform this particular MMR function will result in a higher spontaneous germ line mutation rate. It is known that the mammalian germ line mutation frequency is around 1 x 10 ⁸ per bp per generation (Drake et al, (1998) Genetics 148, 1667-86). By high-throughput resequencing of preselected amplicons, followed by heterozygous mutation discovery in DNA samples from the offspring of 2 homozygous males and wild- type females, the spontaneous single base pair mutation rate in the male germ line of the Msh67- rat was determined. In a total of 8.3 x 10 ⁶ resequenced base pairs 3 mutations were found (twice a C>T and once a G>A transition), indicating that the spontaneous germ line mutation rate is about 3.1 x 10- 7 per bp (Table 2), which is an increase of more than 30-fold. The same amplicons were sequenced in DNA samples from the offspring of 4 wild- type animals (littermates of the Msh67- animals) and no mutation was found in the 6.6 x 10 ⁶ sequenced base pairs.

Msh6 knockout rats show a reduced life span and develop tumors Mouse studies have shown that Msh6 deficiency results in a reduced life span due to the development of different types of tumors (Edelmann et al. (1997) Cell 91, 467-77). We monitored 9 homozygous males and 8 homozygous females and the time they became moribund was recorded. The homozygous mutant rats showed a median survival time of 14 months, whereas 95% of wild type rats normally survive at this age (Figure 3). After 18 months all the Msh6-/- rats had become moribund and there was a tumor incidence of about 88%. These results are consistent with the observation that Msh6- deficient mice show a reduced life span

although they exhibited a median survival time of only 10 months (Table 4).

Each rat was subjected to a complete necropsy and histopathological analysis. Tumor occurrence started at 9 months of age in a number of different locations and the first affected rats were males. Representative histological specimens of the tumor spectra that were observed are shown in figure 4 and the complete data are summarized in table 3. Macroscopical examination revealed a significantly enlarged spleen, which was the result of infiltration of lymphomas. In total 8/17 rats developed highly invasive lymphomas in the spleen, liver, kidney, lung and mediastinum (Figure 4). Histological analysis of the lymphomas revealed that neoplastic lymphocytes were organized in sheets separated by fine fibrovascular stroma. The round cells were uniform and of medium size with scant cytoplasm and round to ovoid nuclei containing fine chromatin and occasionally a prominent nucleolus located centrally (lymphoblastic lymphoma). On average there was one mitotic figure per 40 x high power field. Sheets of neoplastic cells infiltrate into the surrounding tissue and occasionally into vascular structures. Multiple areas within the neoplasm show apoptosis characterized by pyknotic and fragmented nuclei and bright eosinophilic cytoplasm (starry sky pattern). To ascertain the cellular origin of the lymphomas, sections of tumors were stained with CD79, a B-cell specific antibody and a CD3 antibody that is T-cell specific. Of these, 7 were determined to be B-cell lymphoblastic lymphoma and one a T cell lymphoblastic lymphoma. Other tumors found in male mshβ ^ rats included a testicular leydig cell tumor, and a mammary fibroadenoma and adenocarcinoma. Four females suffered from vaginal bleeding and histological evaluation revealed the presence of endometrial carcinomas (Figure 4D) in three animals and uterine leiomyosarcoma in one female (Figure 4C). One female developed a gastric squamous cell carcinoma (Figure 4F). Notably, some of the tumor reached big sizes, like some of the lymphomas that had an estimated diameter in excess of 3 cm. Remarkably, the leiomyosarcoma in the uterus

reached a diameter of ±5 cm and also the gastric tumor reached a big size with a diameter of approximately 2 cm (Table 3).

Example 2

Materials and Methods Animals and ENU mutagenesis protocol

All experiments were approved by the Animal Care Committee of the Royal Dutch Academy of Science according to the Dutch legal ethical guidelines. Experiments were designed to minimize the number of required animals and their suffering. Male MSH6 knockout rats (Msh6 ^1Hubr ) of 12 weeks of age were given three weekly intraperitoneal injections of 25, 30 and 35 mg/kg bodyweight ENU. Preparation of ENU (Isopac; Sigma) was done as decribed (for detailed description, see (Smits et al. (2006) Pharmacogenet Genomics 16, 159-69) After the treatment the injected males were monitored for fertility by breeding with one or two females. Pups from these early matings were counted, but not analyzed. Ten weeks after the last injection, fertile males of the highest-dosed fertile group were kept on a weekly breeding scheme with two females to produce Fl progeny for mutational analysis.

Animals were housed under standard conditions in groups of two to three per cage per gender under controlled experimental conditions (12-h light/dark cycle, 21±1°C, 60% relative humidity, food and water ad libitum).

Genomic DNA isolation

At two weeks of age Fl progeny were uniquely tagged by fetching ear clips. Ear clips were lysed for at least 2 hours at 55°C and 1400 rpm in a Thermomixer comfort (Eppendorf) in 400 μl lysisbuffer (100 mM Tris (pH 8.5), 200 mM NaCl, 0.2% SDS, 5 mM EDTA and 100 μg/ml of freshly added Proteinase K) followed by phenol/chloroform (1:1, vol/vol) extraction. The DNA was precipitated by adding 300 μl isopropanol, mixing and centrifuging for 20 min, at 21,000 g at 4°C. The supernatant was discarded and pellets were washed with 70% ethanol and dissolved in 500 μl MQ.

PCR and sequencing conditions

768 pre-selected amplicons were amplified using a nested PCR setup. The first PCR was carried out using a touchdown thermocycling program (94°C for 60 sec; 15 cycles of 92°C for 30 sec, 65°C for 30 sec with a decrement of 0.6 ⁰ C per cycle, 72°C for 60 sec; followed by 30 cycles of 92°C for 30 sec, 58°C for 30 sec and 72°C for 60 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems). The PCR reaction contained ±100 ng DNA, 0.2 μM of each forward primer and 0.2 μM of each reverse primer, 400 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% dimethyl sulfoxide (DMSO) (w/v), 2 mM MgCk, 85 mM ammonium acetate pH 8.7 and 0.2 U Taq Polymerase in a total volume of 5 μl. After thermocycling, the PCRl reactions were diluted with 20 μl H2O and 1 μl was used as template for the second round of PCR. The second PCR was done using a standard thermocycling program (94°C for 60 sec; 35 cycles of 92°C for 20 sec, 58°C for 30 sec and 72°C for 60 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems). PCR2 mixes contained 1 μl diluted PCRl template, 0.1 μM forward primer, 0.1 μM reverse primer, 100 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% DMSO (w/v), 2 mM MgCl ₂ , 85 mM ammonium acetate pH 8.7 and 0.1 U Taq Polymerase in a total volume of 5 μl. Several samples of each plate were tested on a 1% agarose gel containing ethidium bromide for the presence of the proper PCR fragment. PCR2 products were diluted with 20 μl H ₂ O and 1 μl was directly used as a template for the sequencing reactions. Sequencing reactions contained 0.1 μl BigDye (v3.1; Applied Biosystems), 1.9 μl BigDye Dilution Buffer (Applied Biosystems) and 0.4 μM of the forward primers used for the

PCR2 reaction in a total volume of 5 μl. The thermocycling program that was recommended by the manufacturer was used for the sequence reactions (40 cycles of 94°C for 10 sec, 50 ⁰ C for 5 sec and 6O ⁰ C for 120 sec). Sequencing products were purified by ethanol precipitation in the presence of 40 mM sodium-acetate and analyzed on a 96-capillary 3730XL DNA analyzer (Applied Biosystems), using the standard RapidSeq protocol on 36 cm array. Sequences were analyzed for the presence of heterozygous

mutations using PolyPhred (Nickerson et al. (1997) Nucleic Acids Res. 25, 2745-51) and in-house developed software. All candidate mutations were verified in independent PCR and sequencing reactions.

Project management and primer design using LIMSTILL

All resequencing projects were managed using LIMSTILL, LIMS for Induced Mutations by Sequencing and TILLing (V.G., E. C, unpublished). This web-based publicly accessible information system (http://limstill.niob.knaw.nl) was used to generate projects and visualize gene structures based on Ensembl genome data, the design of PCR primers, entry, archiving and primary interpretation of mutations. The primer design application within LIMSTILL is Primer3-based and parameters are set to design primers with an optimal melting temperature of 58°C. Predictions on the effect of amino acid changing mutations were also conducted within LIMSTILL and are based on stand-alone versions of two prediction programs: SIFT (Ng et al. (2003) Nucleic Acids Res. 31, 3812-14) and PolyPhen (Ramenskyet al. (2002) Nucleic Acids Res. 30, 3894-900). Calculations of introducing a premature stop codon with different mutation spectra are integrated into LIMSTILL and done for all 768 working amplicons.

Phenotyping of mutant animals

At 60 days of age animals were phenotyped in an open field for 30 minutes using the Phenotyper® (Noldus; the Netherlands) followed by a modified SHIRPA protocol (http://empress.har.mrc.ac.uk/). Subsequently, animal behavior, food and water intake were observed for 48 hours with the Phenotyper®. A range of phenotypes were calculated from the animal tracking data, including activity in various time bins, distance moved, speed of movement, time spent in central area, time spent in shelter, total food in take, total water intake. Aberrant phenotypes were scored as compared to the average of other mutants and wild type animals (Table 9).

Results

Effect of ENU on MSH6 knockout rat

The efficiency of ENU is dependent of the dose, reflected by a higher mutagenicity after three weekly administrations of low doses of the mutagen compared to a single high-dose injection (Justice et al. (2000) Mamm Genome 11, 484-8). Furthermore it has been shown that higher ENU concentrations increase the mutation frequency (Smits et al. (2004) Pharmacogenet Genomics 16, 159-69; Smits and Cuppen (2006) Trends Genet. 22, 232-40). To determine the optimal ENU concentration we performed an ENU dose-response experiment and determined the fertility of the founders. At 12 weeks of age mshβ ¹ ' males were treated with 3 weekly doses of different concentrations of the mutagen. Fertility was determined after a full cycle of spermatogenesis (± 10 weeks) by breeding and/or histological analysis of the testes. At concentrations of 35, 30 and 25 mg/kg, respectively 1/8, 5/8 and 4/6 males were fertile (Table 5).

Compared to earlier studies, where at 40 and 35 mg/kg ENU respectively 6/10 and 10/10 rats are fertile (Smits and Cuppen (2006) Trends Genet. 22, 232-40), this is a considerable decrease. It is known that MSH6 deficiency reduces lifespan in both rats (van Boxtel et al. (2007) submitted) and mice (Edelmann (1997) Cell 91, 467-77) as a result of tumor development. MSH6 deficiency increased ENU sensitivity not only in fertility, but also in survival. Whereas untreated mshβ' ¹ ' rats show a median survival of 14 months, this is reduced in the ENU treated males to an average mean of 37 ± 3 weeks (~ 8 months) and no male was alive after 50 weeks of age (Figure 5). This decrease in lifespan was due to the development of tumors, mainly lymphomas (data not shown). In previous studies, ENU treated wild-type males didn't show any reduced lifespan till 1,5 years of age in both 40 mg/kg bodyweight (10 out of 10 survived) and 35 mg/kg bodyweight (9 out of 10 survived) ENU treated males (unpublished data).

Mutation resequencing protocol

Recently we have developed an automated, high-throughput and very cost- effective resequencing protocol using dideoxy sequencing (Smits et al. (2006) Pharmacogenet Genomics 16, 159-69). To reduce the costs even more, we designed a screening scheme, whereby the offspring of the mutagenized animals can be screened before weaning. As a result less animal-room space is needed increasing the cost-efficiency of the method. At two weeks of age tissue samples from Fl offspring was taken and DNA was isolated (Figure 6A). Every pup was screened for 768 pre-selected amplicons by PCR amplifying and sequencing using two 384 wells plates that contained unique primers in every well (Figure 6B). The sequencing files of the same amplicons for different pups were aligned, processed using PolyPhred (Nickerson et al. (1997) Nucleic Acids Res. 25, 2745-51) and inspected with in house developed software (Figure 6C). Mutations were verified and interesting mutants were weaned (Figure 6D). This procedure was repeated for all new Fl offspring.

Increased ENU mutagenicity in MSH6 deficient background

The optimal ENU treatment in MSH6 deficient background was 3 x 30 mg/kg bodyweight ENU in terms of fertility after a full cycle of spermatogenesis. It has been shown that in wild-type rats the highest possible dose, in terms of fertility, gave increased mutation frequencies (Smits et al. (2004) Genomics 83, 332-4; Smits et al. (2006) Pharmacogenet Genomics 25, 2745-51). After screening 291 Fl progeny of this group, covering almost 70 Mb, we discovered 96 ENU-induced mutations (Table 5). This results in mutation rate of 1 mutation every 7.26 x 10 ⁵ base pairs, a 1.7x increase compared to control animals.

Only one nest of three pups could be recovered from the only fertile male treated with 3 x 35 mg/kg bodyweight ENU. In the 0.72 Mb we sequenced, one mutation was discovered resulting in a comparable mutation rate as the 3 x 30 mg/kg treated group (Table 5). However the number is too low to conclude anything from this group. One male treated with 25 mg/kg

bodyw eight ENU produced 16 progeny that were screened. In this group 3.6 Mb was sequenced and 5 mutations were discovered, again resulting in a comparable mutation rate as observed in the 3 x 30 mg/kg treated group (Table 5).

ENU-induced mutation frequency in MSH6 deficient background differs in time

Besides increasing the cost-efficiency, screening the Fl progeny before weaning also allows for mutation discovery in time. Offspring from the males treated with 3 x 30 mg/kg bodyweight ENU show a reduction of mutation frequency in time (Table 6). In the 59 animals that were born 14 - 16 weeks after the ENU injection we sequenced ± 15 Mb and found 28 mutations, resulting in a mutation frequency of 1.88 x 10 ⁶ per base pair and a rate of 1 mutation every 5.3 x 10 ⁵ base pairs. Animals that were born 17 - 19 weeks and 20 - 22 weeks after the last ENU injection show a decrease in mutation frequency of respectively 1.65 x 10 ⁶ and 1.24 x 10 ⁶ per base pair. Fl progeny that was born 23 weeks after the last injection or later show a steady frequency of approximately 1.0 x 10 ⁶ per base pair. These last observations are accompanied by the high mortality rate in the founders (average mean of survival was 37 ± 3 weeks, which is 23 weeks after the last injection, Figure 5). Remarkably, all fertile 3 x 30 mg/kg bodyweight treated ENU males were still alive at this point. It is tempting to speculate whether there is a relationship between fertility and viability in these ENU treated MSH6 deficient males. However, the numbers of treated males are too low to make any conclusions about this remark.

ENU-induced mutations and spectrum

In order to maximize the outcome of interesting mutants we mostly resequenced large exons of G protein-coupled receptors (GPCRs). It has been estimated that approximately 80% of all known hormones and neurotransmitters signal via GPCRs and that 30-45% of the current drug target these receptors (Gloriam et al. (2007) BMC Genomics 8, 338). Of the

102 mutations we have found, 94 are in coding regions (Table 7). Four mutations introduce a premature stop codon in the open reading frame and most likely will result in a functional knockout of that gene (nonsense, ± 4%), 70 mutations cause an amino acid change (missense, ± 74%) and 20 mutations do not effect protein coding (silent, ± 21%).

Three of the four new knockouts represent orphan GPCRs. In humans GPR19 encodes a gene that demonstrated closest similarity to D2 dopamine and neuropeptide Y receptors. Furthermore GPR19 expression significantly overlapped D2 dopamine receptor gene expression in peripheral and brain regions (O'Dowd et al. (1996) FEBS Lett. 394, 325-9). Nothing much is know about GPR84 except that in humans is expressed in brain, heart, muscle, colon, thymus, spleen, kidney, liver, intestine, placenta, lung and leukocytes (Wittenberger et al. (2001) J MoI Biol. 307, 799-813). PSYR was designated T-cell death-associated gene-8 (Tdag8) in mice and demonstrated an elevated expression upon activation-induced apoptosis of T-cell hybridomas (Choi et al. (1996) Cell Immunol 168, 78- 94). Finally CXCR2 encodes the interleukin 8 receptor, type 2 and was shown to be a major mediator of neutrophil migration to sites of inflammation (Cacalano et al. (1994) Science 265, 682-4). As in our previous studies (Smits et al, (2006) Pharmacogenet Genomics 16, 159-69), we used PolyPhen (Ramensky et al. (2002) Nucleic Acids Res. 30, 3894- 900) and SIFT (Ng et al, (2003) Nucleic Acids Res. 31, 3812-4) software to predict the potential effect of the amino acid changes resulting from the missense mutations. Of the 70 missense mutations 30 were predicted to affect protein function by both programs (Table 3). Out of the remaining 40 missense mutations 30 were predicted to affect protein function by one of the two programs and 10 were predicted not to affect protein function by both programs (Table 7). Remarkably, ENU-induces a different mutation spectrum under MSH6 deficient compared to MMR proficient background (Figure 7A). There is a significant difference between the spreads of the transitions under MSH6 proficient and deficient backgrounds (p = 0.028), but not between the

transversions. The difference in mutation spectra most probably reflects MutSα mismatch binding preference. In Lad mutational reporter studies it was shown that MSH6 deficient small intestinal epithelial cells of mice had a predominance of spontaneous G:C to A:T transitions (Mark et al. (2002) Oncogene 21, 7126-30). Furthermore mshϊ'- mouse ES cell treated with ENU show predominance for A:T to T:A transversions and G:C to A:T transitions (Claij et al, (2003) Cancer Res. 63, 2062-6). This change in mutation spectrum therefore underlines the specific effect ENU in MSH6 deficient background. The chance of introducing a stop codon depends on the mutation spectrum. This chance is unequally divided between the different mutation types. For example, G: C to A:T transitions can introduce 5 premature stop codons out of the 183 total changes in all possible codons, whereas A:T to G:C transitions will never lead to a premature stop codon. To determine the difference in chance of introducing a premature stop codon we calculated this chance for the 768 amplicons of the resequencing panel with the wild-type versus the mutant mutation spectra. The chance of introducing a premature stop codon with the wild-type mutation spectrum is ~ 4.3%, whereas the chance using the mshβ ¹ ' mutation spectrum is significantly (p « 0.01) elevated to ~ 5.0% (Figure 7B). This means a 16% higher chance of introducing a premature stop under MSH6 deficient background compared to MSH6 proficient background. The combined effect of the increased ENU-induced mutation frequency and change in mutation spectrum in the msh6-deficient rat results in an about 2.5-fold increase in the chance of identifying a knockout allele for any given gene (Figure 8). In line with this, the total number of animals that would be needed to identify knockouts for 95% of all known rat genes decreases from 110,000 to about 40,000 when applying ENU-driven target- selected mutagenesis in a msh6-deficient genetic background. Similarly, when considering a collection of 5,000 Fl rats derived from ENU treated mshδ-deficient rats, knockout alleles for ~50% of all the rat genes will be

present, while this would only be 20% in a wild-type background (Figure 8).

In support of the bioinformatics-based prioritization, we analysed the in vivo effects of the mutations by crossing the mutations to homozygosity. The observed phenotypes can be found in Table 9.

Example 3

Materials and Methods

Animals and ENU mutagenesis protocol

All experiments were approved by the Animal Care Committee of the Royal Dutch Academy of Science according to the Dutch legal ethical guidelines. Experiments were designed to minimize the number of required animals and their suffering. Male MSH6 knockout rats (Msh6 ^1Hubr ) and heterozygous male littermates of 12 weeks of age were given three weekly intraperitoneal injections of 30 mg/kg body weight ENU. Preparation of ENU (Isopac; Sigma) was done as decribed (for detailed description, see (Smits et al. (2006) Pharmacogenet Genomics 16, 159-69) After treatment, the injected males were monitored for fertility by breeding with one or two females. Pups from these early matings were counted, but not analyzed. Ten weeks after the last injection, fertile males of the highest-dosed fertile group were kept on a weekly breeding scheme with two females to produce Fl progeny for mutational analysis. Animals were housed under standard conditions in groups of two to three per cage per gender under controlled experimental conditions (12-h light/dark cycle, 21±1°C, 60% relative humidity, food and water ad libitum).

Genomic DNA isolation At two weeks of age Fl progeny were uniquely tagged by fetching ear clips. Ear clips were lysed overnight at 55°C and 1000 rpm in a Thermomixer comfort (Eppendorf) in 400 μl lysisbuffer (100 mM Tris (pH 8.5), 200 mM NaCl, 0.2% SDS, 5 mM EDTA and 100 μg/ml of freshly added Proteinase K) followed by two phenol/chloroform (1:1, vol/vol) extractions. The DNA was precipitated by adding 300 μl isopropanol, mixing and centrifuging for 20 min, at 21,000 g at 4°C. The supernatant

was discarded and pellets were washed with 70% ethanol and dissolved in 500 μl MQ.

PCR and sequencing conditions 768 pre-selected amplicons were amplified using a nested PCR setup. The first PCR was carried out using a touchdown thermocycling program (94°C for 60 sec; 15 cycles of 92°C for 30 sec, 65°C for 30 sec with a decrement of 0.6 ⁰ C per cycle, 72°C for 60 sec; followed by 30 cycles of 92°C for 30 sec, 58°C for 30 sec and 72°C for 60 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems). The PCR reaction contained ±100 ng DNA, 0.2 μM of each forward primer and 0.2 μM of each reverse primer, 400 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% dimethyl sulfoxide (DMSO) (w/v), 2 mM MgCl ₂ , 85 mM ammonium acetate pH 8.7 and 0.2 U Taq Polymerase in a total volume of 10 μl. After thermocycling, the PCRl reactions were diluted with 20 μl H ₂ O and 1 μl was used as template for the second round of PCR. The second PCR was done using a standard thermocycling program (94°C for 60 sec; 35 cycles of 92°C for 20 sec, 58°C for 30 sec and 72°C for 60 sec; 72°C for 180 sec; GeneAmp9700, Applied Biosystems). PCR2 mixes contained 1 μl diluted PCRl template, 0.2 μM forward primer, 0.2 μM reverse primer, 100 μM of each dNTP, 25 mM tricine, 7.0% glycerol (w/v), 1.6% DMSO (w/v), 2 mM MgCl ₂ , 85 mM ammonium acetate pH 8.7 and 0.1 U Taq Polymerase in a total volume of 10 μl. Several samples of each plate were tested on a 1% agarose gel containing ethidium bromide for the presence of the proper PCR fragment. PCR2 products were diluted with 20 μl H ₂ O and 1 μl was directly used as a template for the sequencing reactions. Sequencing reactions contained 0.1 μl BigDye (v3.1; Applied Biosystems), 3.8 μl BigDye Dilution Buffer (Applied Biosystems) and 0.4 μM of the forward primers used for the PCR2 reaction in a total volume of 10 μl. The thermocycling program that was recommended by the manufacturer was used for the sequence reactions (40 cycles of 94°C for 10 sec, 50 ⁰ C for 5 sec and 60 ⁰ C for 120 sec). Sequencing products were purified by ethanol precipitation in the

presence of 40 niM sodium-acetate and analyzed on a 96-capillary 3730XL DNA analyzer (Applied Biosystems), using the standard RapidSeq protocol on 36 cm array. Sequences were analyzed for the presence of heterozygous mutations using PolyPhred (Nickerson et al. (1997) Nucleic Acids Res. 25, 2745-51) and in-house developed software. All candidate mutations were verified in independent PCR and sequencing reactions.

Biochemical analyses of knockout animals

For western blot analysis proteins were extracted by adding 300 μl of RIPA lysis buffer (150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl pH 8.0, 1 mM PMSF and 1.0 mM sodium ortho-vanadate) to approximately 0.25 g of tissue. The tissues were homogenized and then maintained under constant agitation for 2 hours at 4°C. The lysates were spinned for 20 minutes at 14000 rpm at 4°C and the supernatant was transferred to a new tube. One volume of SDS loading buffer (125 mM TRIS-CL pH 6.8, 4% SDS, 20% glycerol, 100 mM freshly added DTT and 0.02% bromophenol blue) to one volume of lysate and then incubated for 10 min at 70 ⁰ C. The protein was separated on a SDS gel (15% acrylamide gradient, Bio-Rad) and transferred to a nitrocellulose membrane. Non-specific, protein-binding sites were blocked by incubating the membrane with blotting buffer: 0.05% PBS-Tween containing 5% nonfat dry milk for 3 hour at room temperature. The membrane was washed with 0.05% PBS-Tween and incubated overnight at 4°C with a 1:500 dilution of a polyclonal rabbit anti-mouse I18rb antibody (Santa Cruz Biotechnology) in blotting buffer. The membrane was washed five times for 10 min in 0.05% PBS-Tween and then incubated for 2 hours with peroxidase-conjugated, anti-rabbit IgG (Chemicon) diluted 1:2000 in blotting buffer at room temperature. The membrane was washed five times with 0.05% PBS-Tween and protein bands were detected by using the enhanced chemiluminescence detection method

(ECL, Amersham Biosciences). Liver lysate from rat was used as a control and total protein was measured by Coomassie Brilliant Blue staining.

Immunohistochemistry was performed by fixing tissues 0/N with phosphate buffered 4% formaldehyde at 4°C. Slides were blocked for 1 hour at RT with 1% BSA in PBS, followed by O/N incubation at 4°C with 1:200 dilution of a polyclonal rabbit anti-human Gpr65 antibody (Abeam) in blocking buffer. Slides were washed 3 times 10 minutes at RT with PBS and incubated with System HRP polymer-conjugated anti-rabbit for 1 hour at RT. Slides were washed 3 times 10 minutes with PBS and incubated for 10 minutes with DAP at RT. Slides were washed 3 times 10 minutes with PBS, dehydrated and covered with pertex.

Archiving of rat sperm

Male rats of at least 12 weeks of age were sacrificed and the sperm was cryopreserved as described (Mashimo et al., (2008) Nat. Genet. 40, 514-5). Briefly, epididymes were removed and dispersed in 2 ml of primary freezing medium and incubated for 3 minutes at room temperature, for 30 minutes at 20 ⁰ C and for 30 minutes at 4°C. 2 ml of cold 2 x secondary freezing medium was added and samples were sucked into a straw and placed 2 cm above liquid nitrogen for 15 minutes. Finally, the straws were put into liquid nitrogen and archived.

Results

Increased sensitivity and ENU mutagenesis efficiency in MMR-deficient background

Male Mshβ '- and Msh6 ^+/ - were treated with 30 mg/kg bodyweight ENU and after a full cycle of spermatogenesis used for generating mutant Fl progeny. Both survival and fertility of the Mshβ ^ founders was affected by the ENU treatment as shown before (van Boxtel et al., (2008) BMC Genomics, 9:460) while the Msh6 ^+/ - founders are less affected (Table 10). We screened 375 Fl progeny of Mshβ ^ founders that were born 14 — 29 weeks after the last injection, covering more than 90 Mb and discovered 128 mutation, resulting in a mutation rate of 1 in 7 x 10 ⁵ bp (Table 10), which is comparable with earlier findings (van Boxtel et al., (2008) BMC

Genomics, 9:460). Furthermore, we screened 37 Fl animals of heterozygous founders that were born 14 - 18 weeks after the last injection and discovered 7 mutations in almost 10 Mb resequenced bp, resulting in a mutation rate of 1 in 1.42 x 10 ⁶ bp (Table 10). These findings further underline the increased efficiency of ENU mutagenesis in a MMR- deficient background.

ENU-induced mutations

For high-throughput mutation discovery we resequenced a panel of 768 amplicons, which mainly contains one-to-one orthologues of human G protein-coupled receptors (GPCR) and was used in earlier studies (van Boxtel et al., (2008) BMC Genomics, 9:460) with some changes. Of the 135 mutations that were identified in the Fl progeny of both Msh6 ^+/ - and Msh6''' founders 117 resided in coding regions. Six of the mutations result in the introduction of a premature stopcodon (nonsense, ± 5.1%) and most likely represent functional knockout of the genes (Table 10). Furthermore, 86 mutations result in a amino acids change (missense, ± 73.5%) and 25 did not result in amino acid changes (silent, ± 21.4%). Five of the 6 nonsense mutations reside in GPCR genes, of which 2 are orphan receptors, Gpr63 and Gpr85 (Table 11). The latter is also known as super conserved receptor expressed in brain (SREB), which is characterized by its extremely high mammalian conservation and was shown to be involved in determining brain size and potentially in vulnerability to schizophrenia (Matsumoto et al., (2008) Proc. Natl. Acad. Sci. 105, 6133-8). Ccr4 is a chemokine receptor, which was shown to be a key regulator of innate immune responses (Ness et al., (2006) J. Immunol. 177, 7531-9). The Melanocortin 4 receptor (MC4R ) has been shown to have an anorexic effect on feeding behavior and disrupting this gene in mice results in obesity (Huszar et al., (1997) Cell 88, 131-41). There is not much known about the serotonin receptor 5HTR1F, which is expressed throughout the brain. The nonsense mutant that was discovered in a non- GPCR gene resides in breast cancer resistant protein (Bcrp), which plays a

major role in the multidrug phenotype of specific human breast cancer (Doyle et al., (1998) Proc. Nat. Acad. Sci. 95, 15665-70). Furthermore, it was shown to transport pheophorbide-a, a chlorophyll-breakdown product. Mice lacking transporter became extremely sensitive to this breakdown product and suffered from severe and sometimes lethal photoxic lesion on light-exposed skin (Jonker et al., (2002) Proc. Nat. Acad. Sci. 99, 15649- 54).

A total of 61 missense mutations were identified in GPCRs genes. Three of these were found in two Fl animals, which always derived from the same Mshβ '- founder and are thus most likely the result of a clonal effect, which is not unlikely with the increased mutation frequency combined with an increased lethality in spermatogonial stem cells. Importantly, 24 of these missense mutations are located in transmembrane (TM) domains (Table 11) and 7 change a hydrophobic residue into a hydrophobic, which can cause incorrect folding and incorporation into the membrane. Notably, the mutation in the orexin2 receptor (OX2R) changes the tryptophan, which is part of the highly conserved (D/E)R(Y/W) motif that has been shown to interact with the G protein and thus most likely will affect proper receptor functioning. Furthermore, a mutation in the TM 7 domain of an orphan receptor Gpr42 introduces a proline residue, which causes a kink and can severely affect receptor functioning. In addition, we identified 26 missense mutations in non-GPCR genes (Table 11).

Archiving rat sperm After outcrossing the Fl animals, males were sacrificed for sperm cryopreservation. In case of female Fl animals, the sperm of male F2 animals was used. Mutants can be revived by intracytoplasmic sperm injection (ICSI) as decribed (Mashimo et al., (2008) Nat. Genet. 40, 514-5). In total the sperm of 240 male Fl animals and 40 male F2 animals was archived. With an average mutation rate of 1 every 750,000 bp this archive would potentially contain approximately 753 nonsense and 11,091

missense mutations, which can be discovered if the whole ORFeome of the rat would be screened.