CA2135313A1 - Methods for producing transgenic non-human animals harboring a yeast artificial chromosome - Google Patents

Abstract

The present invention provides methods and compositions for transferring large transgene polynucleotides and unlinked selectable marker polynucleotides into eukaryotic cells by a novel method designated co-lipofection. The methods and compositions of the invention are used to produce novel transgenic non-human animals harboring large transgenes, such as a transgene comprising a human APP
gene or human immunoglobulin gene.

Description

W094~00569 ~ 3 PCT/U~93/~5873 M~THODS FOR PRODUCIN~ TR~NSGENIC NON-HUMAN ANIMALS
HARBORING A YEAST ARTIFICIAL CHROMOSOME
1 0 ' TECHNICAL FI~rn ::
The invention relates to transgenic non-human animals capab'~ of e~pressing xenogenic polypeptides, ~. :
transgenes use~ ~0 produce such transgenic animals, transgenes capable of expressing xenogenic polypeptides, yeast artificial chromosomes comprising a polynucleotide sequence encoding a human protein such as a human immunoglobulin or amyloid precurso~ protein (~PP), methods and transgenes for transferring large polynucleotide sequences into cells0 and mekhods for co-lipo~ection of discontinuous polynucleotide s~quences into cells.

BACK~ROUND OF THE INVENTION
Transferring exogenous genetic material into cells is the basis f or modern molecular biology . The contirluing development of novel methods f or improving the ef î iciency, specificity, and/or size limitations of the transfer proces has broadened the scope of research and product development by enablin~ the production of polynucleotide clones and recombinant organisms that previously were impractical or im~ossible to construct. Calcium phosphate precipitation, electroporation, lipofection, ballistic transfer~ DEAE-dextran transfection, microinjection, and viral-based transfer methods, among others, have been described for introducing ~or~ign DNA fra~ments into mammalian cells.
The art al~o has de~el~ped yeast arti~icial chrom~some (i'YAC") cloning vec~ors which are capable of propagating large (50 to more than 1000 kilobases) cloned inserts (U.S. Patent 4,889,806) o~ xenogenic DNA. YAC clone WV94/00569 2 i ~ 5 3 1 3 PCT/VS93/05P-~

libraries have been used to identify, map, and propagate large fragments of mammalian genomic DNA. YAC cloning is especially useful for isolating intact genes, particularly large genes having exons spanning several tens of kilobases or more, and genes having distal regulatory elements located tens of kilobases or more upstream or downstream from the exonic sequences. YAC cloning is particularly advantageous for isolating large complex gene loci, such as unrearranged immunoglobulin gene loci, and genes which have been inexactly mapped to an approximate chromosomal region (e.g., a Huntington's chorea gene). Y~C cloning is also well-suited for making vectors for performing targeted homologous recombination in mammalian cells, since YACs allow the cloning of large conti~uous sequences useful as recombinogenic homology regions in homologous targeting vectors. Moreo~er, YACs afford a system for doing targeted homologous recombination in a yeast host cell to create novel, large : transgenes (e~g., large minigenes, tandem gene arrays, etc.) in YAC constructs which could then be ~ran ferred to mammalian host cells.
Unfortunately, manipulation of large polynucleotides is problematic. Large polynucleotides are susceptible to breakage by shearin~ forces and form highly viscous solutions even at relatively dilute concentrations, making in vitro manipulation exceedingly difficult. For these reasons, and : others, it is desirable to reduce the amount of manipulation that YAC clones and other large DNA fragments ara sub~ected to in the process o~ constructing large transgene constructs or homologous recombination constructs.
More problematic is the fact that the transfer of large, intact polynucleotides into mammalian cells is typically ine ficient or provides a restriction on the siz~ of the polynucleotide transferred. ~or example, Schedl et al.
~ (1992) Nuclei~ ACids Res! 20: 3073, describe transferring a 35 kilobase YAC clone into the mouse genome by pronuclear injec~ion of murine embryos; however, the shear forces produced in the injection micropipette will almost certainly preclude the efficient transfer of significantly larger YAC

~ ~ ~ J ~
W094/00569 P~T/VS93/05873 clones in an intact form. Many large genes likely could not be transferred efficiently into mammalian cells by current microin,~ction methods~
Spheroplast fusion has b~en used to introduce YAC
DNA into fibroblasts, embryonal carcinoma cells, and CHO cells (Pachnie et al. (1990) Proc. Natl. Acad. Sci. (U.S.A.~ 87:
5109; Payan et al. (19~0) Mol. Cell. Biol. lO: 4163; Chirke et al~ (1991) E~aQ_~ 10: 1629; Davies et al. (1992) Nucleic Acids Res. 20: 2693)~ Alternative transfection methods such as calcium phosphate precipitation and lipofection have been used to transfer YAC DNA into mammalian cells ~Eticciri et al.
(1991) Proc. Natl. Acad. Sci. fU.S~A ) 88: 2179; Strauss W and .Jaenisch R (1992) EMBO J. 11: 417).
Thus, there exists a need in the art for an efficient method for transferring large seyments of DNA, such as large YAC clones, into mammalian cells, such as embryonic stem ~ells for making transgenic animals, with a minimum of manipulation and cloning procedures. In partlcular, it would be highly advantageous if it were possible to isolate a large cloned mammalian genomic fragment from a YAC library, either linked to YAC yeast sequenoes or purified away from YAC yeast sequences, and transfer it intact into a mammalian host cell (e.g.5 an ES cell~ with a second polynucleotide sequence (e.g., a selectable marker such as a neo~ expression cassette) without additional cloning or manipulation ~e~g., ligation of the sequences ~o each other). Such a method would allow the efficient construction of transgenic cells, transgenic animals, and homologously targeted cells and ani~als. These ~:
transgenic/homologously targeted cells and animals could provide useful models of, for example, human genetic diseases such as Huntington's chorea and Alzheimers disease, among others.
Alzheimer's Disease ~ At present there is no known therapy for the various forms of Alzheimer's disease (AD)~ H9wever, th~r~ are several disea~e states for which effective treat~ent i5 a~ailable and which give rise to progressive intellectual dete~ioration W094/00569 % 1 ~ 5 3 ~ ., P~T/US93/05~

closely resembling the dementia associated with Alzheimer's disease.
Alzheimer's diseas~ is a progr~ssive disease known gen rally as senile dementia. Broadly speaking the disease falls into two categories, namely late onset and early onset.
Late onset, which occurs in old age (65 + years), may be caused by the natural atrophy of the brain occurring at a faster rate and t~ a more severe degree than normal. Early onse~ Alzheimer's disease is much more infrequent but shows a pathologically identical dementia with brain-atrophy which develops well before the senile period, i.e., between the ages of 35 and 60 years. There is evidence that one form of this . type o~ Alzheimer's disease is inherited and is therefore known as familial Alzheimer's disease (FAD).
In both types of Alzheimer's disease the pathology is the same but the abnormalities tend to be more se~ere and more widespread in cases beginning at an earlier agP- The disease i5 characterized by four types of lesions in the : brain, thes2 are: amyloid plaques around neurons (senile : 20 plaques), amyloid deposits around cerebral blood vessels, neurofibrillary tangle-~ inside neurons, and neuronal cell death. Senile plaques are areas of disorganized neuropil up to 150~m across with extracellular amyloid deposits at the center. C~rebrovascular amyloid deposits are amyloid material surrounding cerebral blood vessels. Neurofibrillary tangles are intracellular deposits of amyloid protein consisting of : two filament-~ twisted about each other in pairs.
The major protein subunit, amyloid ~ protein, is : found in amyloid filaments of both the neurofibrillary tangle ~
and the senile plaque and is a highly aggregating small :
polypeptide of approximate relati~e molecular mass 4,000.
This protein is a cleavage product of a much larger precursvr protein called amyloid precursor protein (APP).
Th~ APP gene is known to be located on human chromo~ome 21~ A locu~ segregating with familial Alzheimer's dis~ase has been mapped to ohromosome 21 ~St. George Hyslop et al (1987) Science 235: 885) close to the APP gene.
Recombinants between the APP gene and the AD locus have been ~-13~3~ ~
WO9~/0~69 PCT/US93/05873 previously reported (Schellenberg et al. (1988] Science 241:
1507; Schellenberg et al. (1991) Am. J._Hum. Genetics 48: 563;
Srhellenberg et al. (19gl~ Am. J. Hum. Genetics 49: 511, incorporated herein by reference). The development of experimental models of Alzheimer's disease that can be used to define further the underlying biochemical events involved in AD pathogenesis would be highly desirable. Such models could presumably be employed~ in one application, to screen for agents that alter the degeneratiYe course of Alzheimer's disease. For example, a model system of Alzheimer's disease could be used to screen for environmental factors that induce or accelerate the pathogenesis of AD. In contradistinc~ion, . an experimental model could be used to screen for agents that inhibit, pre~ent, or r2verse the progression of AD.
Presumably, such models could be employed to develop pharmaceuticals that are effective i~ preventing, arrestin~
or reversing AD~
Unfortunately, only humans and aged non-human primates develop any of the pathological features of AD; the expense and difficulty of using primates and the length of time required for developin~ the AD pathology makes extensive research on such animals prohibiti~e. Rodents do not develop AD, even at an extreme age. It has been reported that the injection of ~-amyloid protein (~AP) or cytotoxic ~AP
fragments into rodent brain results in cell loss and induces an antigenic marker for neurofibrillary tangle components (Kowall et al. (19913 Proc. Natl._ Acad. Sci. (U.S~A.~ 88:
7247 ) . Mice which carry an extra copy of the APP gene as a r~sult of partial trisomy of chromosome 1~ die before birth (Coyle et al. (1988~ Trends in Neurosci. 11: 390~. Since the cloning of the APP gene, there have been several attempts to produce a mouse model for AD using transgenes that include all or part of the APP gene, ur~f ortunately much of the work remains unpublished since the mice were nonviable or f ailed to 3 5 show AD-like patholo~y; two publish~d reports were retract~d because o~ irregularities in reported results (2~arx J Science 255: 120~), W094J0~s69 ~ 3 ~ ~ PCT/~S93/05~-~

Thus, there is also a need in the art for transgenic nonhuman animals harboring an intact human APP
gene, either a wild-typP allele, a disease-associated allele, or a combination of these, or a mutated rodent ~e.g., murine) allele which comprises sequence modifications which correspond to a human ~PP sequence. Cell strains and cell lines (e.g., astroslial cells) derived from such transgenic animals would also find wide application in the art as experimental mod~ls for developing AD therapeutics.
Nonhuman Transqenic Animals Exressinq Human Immunoqlobulin Making human monoclonal antibodies that bind pred~termined antigens is difficult, requiring a source of . viable lymphocytes from a human that has b~en immunized with an antigen of choice and which has made a substantial humoral immune response to the immunogen. In particular, humans general~y are incapable of making a substantial antibody response to a chall~nge with a human (sel~ antigen, un~ortunately, many such human antigens are promising targets for therapeutic strategies involving h-uman monoclonal ~0 antibodies. On~ approach to making human antibodi e3 that are -~
specifically reactive with predetermined human antigens involves producing transgenic mice harboring unrearranged human i ~ unoglobulin transgenes and having functionally disrupted endogenous immunoglobulin gene~s) (Lonberg and Xay, WO 92~03918, Kucherlapati and Jakobovits, WO 91J10741). ~:~
However, efficiently transferring large DNA segments, such as those spanning significant port}ons of a human light or heavy chain immunoglobulin gene locus, presents a potential o~sta le and/or reduces the efficiency of the process of generating th~
transgenic animals.
Based on the foregoing, it is clear that a need exists for nonhuman cells and nonhuman animals h~rboring one or more large, intact transgenes, particularly a human APP
ge~e or ~ human immunogl9bulin transgene(s~. Thus, it is an object of ~he invention her~in to provide method and compositions ~or transfPrring large transgenes and large homologous recombination constructs, usually cloned as YACs, into mammalian cells, especially into embr~onic stem cells.

~l~ 3~
W094/~0s69 PCT/US93/05873 It is ~so an object of the invention to provide transgenic nonhuman cells and transgenic nonhuman animals harboring one or more APP transgenes of the invention.
The references discussed herein are provided solely for their disclosure prior to the filing date of the present application. Nothing herein is to be construed as an admission that the inventvrs arP not entitled to antedate such disclosure by virtue of prior invention. All citations are incorporated herein by reference.
SUMMARY OF THE INVENTION
In accordance with the foregoing objects, in one .aspect of thP invention methods for transferring larye transgenes and large homologous targeting constructs, typicall.y propagated as YACs and preferably spanning at least one complete transcriptional complex, into mammalian cells, such as ES cells, are provided. In one asp~ct, the methods pro~ide for transferring the large transgenes and large homologous targeting constructs by a lipofection method, such as ~o-lipofection, wherein a second unlinked polynucleotide is transferred into ~he mammalian cells along with the large transgene and/or large homolog~u targeting construct.
Preferably, the s~cond po~ynucleotide confers a s~lectable phenotype (e.g., resistance to G418 selecti~n) to cells which Aave taken up and integrated the polynucleotide sequence~s).
Usually, the large transgene or homologous targeting construct i5 transferred with yeast-derived YAC sequences in polynucleotide linkage, but yeast-derived YAC sequences may be removed by restriction enzyme digestion and separation (~.g, ::
pulsed gel elec~rophoresis). TAe l~rge tran~gene(s) and/or homologous targeting construct~s~ are generally mixed with the ~-~unlinked second polynucleotide (e.g., a neoR expression cassett~ to confer a selectable phenotype) and contacted with a cationic lipid (e.g., DOGS, DOTMA, DOTAP) to form cationic lipid-DNA complexe~ which are ~-ontacted with mammalian cells (e.gO, ES cells) in conditions suitable for uptake o~ the DNA
into the cells (e.g., cul~ure medium, physiological phosphate bu~fered saline, serumfree ES medium). Generally, cells WO9~/0~69 ~ 3~ 3 PCT/US93/oS~

harboring the large transgene or large homologous targeting construct concomitantly harbor at least one copy of the second polynucleotide, so that selection for cells harboring the second polynu lPotide have a significant probability of also harboring at least one copy of the large transgene or large homologous targeting construct, generally as an integrated or homologously recombined segment of an endogenous chromosomal locus. Hence selection for the second polynucleotide (e.g., neoR expression cassette) generally also selects cells harboxing the large transgene or large homo~ogo~s targeting construct without requiring cumbersome polynucleotide linkage (i.e., ligation) of the large transgene or large homologous .targeting construct to the second polynucleotide prior to lipofection. According to the co-lipofection methods of the invention, large segments of xenogenic DNA are rapidly and efficiently transferred into mammalian cells (e.g., murine ~S
cells) without requiring linkage of a selectable marker gene and subsequent cloning.
The invention also provides mammalian cells, preferably ES cells, harboring at least one copy of integrated or homologously recombined large xenogenic ~preferably heterologous) mammalian genomic DNA sequences linked to yeast-derived YAC sequences. Preferably, the large xenogenic (preferably heterologous~ mammalian genomic DNA sequences comprise a complete structural gene, more preferably a - complete trans~.riptional unit, and in one embodiment a complete human APP gene. Typically, the resultant transganic mammalian cells also comprise at least one integrated copy o the unlin~ed second polynucleotide (e.g., the selectable marker), which is usually nonhomologously integrated into at ~east one chromosomal locus, sometimes at a chromosomal locus distinct from that a~ which the large transgene(s) or large homologous targeting construct(s) has been incorporated.
During the transfection process and shortly thereafter, novel mammalian cells, such as ES c~lls, comprising large foreign DNA sequences, an unlinked selectable marker gene, and a suitable cationic lipid are formed, such novel m~mmalian cells are one aspect of the present invention.

W~94~00569 PCT/~S93~05873 The invent~on also provides transgenic nonhuman animals comprising a genome having at least one copy of integrated or homologously recombined large xenogenic (preferably heterologous) mammalian genomic DNA sequences linked to yeast-derived YAC sequences. Preferably, the large xenogenic (preferably heterologous) mammalian genomic DNA
sequences comprise a complete structural gene, more preferably a complete transcrip~ional unit, and in one embodiment a complete human APP gene. Typically, the resultant transgenic nonhuman ma~al also comprises a genome having at least one integrated copy of the unlinked second polynucleotide (e.g., the selectable marker), which is usually nonhomologously integrated into at least one chromosomal locus, sometimes at a chromosomal locus/loci distinct from that at which the large transgene~s) or large homologous targeting construct(s) has/~ave been incorporated. Pre~erably, the large transgene and/or large homologous targeting construct which has been incorporated into a chromosomal locus (or loci) of the nonhuman ani~al is expressed, more preferably is expressed similarly to the naturally-occurring homolog gene in the non-human animal species (e.g., in a similar tissue-specific pattern and/or developmental pattern).
The invention also provides compositions for co-llpofection: transgenesis compositions and homologous targeting compositions for transferring xenogenic, typically heterologous, large (i.e., 50 kb or more) polynucleotide~ into -~
-~ mammalian cells, such as ES cells for making transgenic .;
: nonhuman animals harboring at least one copy of at least one integrated large foreign transgene and/or harboring at least one homologously targeted construct in its genome. A
transgenesis composition comprises: (1) at least one large transgene species, (2) at leas~ one unlinked second polynucleotide species (such as an expression cassette containing the selectable marker gene neoR), and (3) at least one species of suitable cationic lipid. A homologous targ~ting composition co~prises~ at least one large homologous targeting construct species, (2) at l~ast o~e unlinked second polynucleotide species (such as an exp ~ssion W~4/~0569 ~ 3 ~ ~ ~CT/U~93/05~

cassette containing the selectable marker gene neoR), and (3) at least one species of suitable cationic lipid. Preferably the large transgene or large homologous targeting construct spans an entire transcriptional unit. One preferred embodiment of a co-lipofection composition is a composition comprising: (-) a human APP gene sequence ~or a modified murine or rat APP gene having a non-naturally occurring sequence corresponding to a human APP sequence) linked to yeast derived YAC sequences, (2~ an expression cassette encoding a selectable marker, and (3) a suitable cationic lipid. Another preferred embodiment of a co-lipofection composition is a composition comprising: (13 a human . unrearranged immunoglobulin gene sequence (heavy or light chain gene sequence comprising at least two V gene complete segment, at least one complete D segment (if heavy chain gene)~ at least one complete J segment, and at least one constant region gene) linked to yeast-derived YAC sequences, : (2~ an expression cassette encoding a selectable marker, and (3) a suitable cationic lipid.
In one aspect of the invention, multiple species of unlinked polynucleotide sequences are co lipofect~d into murine embryonic stem cells andfor other mammalian cells, wherein at leas~ one species of the unlinked polyn~cleotide sequences ~omprises a selectable marker gene which confers a selectable phenotype to cells which have incorporated it. The resultant cells are selected for the presence of the selecta~le marker; such selected cells have a signif icant probability of romprising at least one integrated copy of the other species of polynucleotide sequence(s) introduced into the cells.

BRIEF DESCRIPTION OF THE FIGUR~S
Fig. 1: Chemical structures of representative cationic lipid~ for forming co-lipofection complexes of the present in~ention.
Fig. 2: PCR analysis of ES clones co-lipof ected with the human APP transgene. Shaded circl s denote wells which were not used . Row pools (A-P~ contained 18 (A-H) or 16 ( ~ P) ~ l ~ 3~
W094/00569 PCT/US93/05$73 clones each. Column pools (P1-P18) contained 16 (P1-P12) or 8 (P13-P18) clones each.
Fig. 3: PCR analysis of ES clones co-lipofected with the human APP transgene. Pools P3, P4, P9, P10, P11, and P12 and pools G, H, K, M, N, 0, and P were candidates for containing clones with both promoter and exon 17 sequences.
Fig. 4: PCR analysis of ES clones co-lipofected with the human ~PP transgene.
Fig. 5: Southern blot analysis of YAC clone DNA
using a human ~lu sequence probe.
Fig. 6: Partial restriction digest mapping of hum~n APP YAC.
Fig. 7: PCR analysis of RNA transcripts expressed from intsgra~ed human APP transgene.
Fig. 8: Quantitative RNase protection assay for detecting APP ~NA txanscripts from human APP transgene.
DefinitionS
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary ~kill in the art to which this inven ion belo~gs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, the preferred methods and materials are described. For purposes of the present invention, the following terms are defined below.
The term "corresponds to" is used herein to mean that a polynucleotide sequence is homologous (i.e., is identical r not strictly evolutionarily related) to all or a portion of a reference polynucleotide se~uence, or that a ., polypeptide sequ~nce is identical to a reference polypep~ide seguence. In contradistinction, the term ~'complementary to"
is used herein to mean that the complemsntary sequence is homologous to all or a portion of a reference polynucleotide sequence. For illustration, the nucleo~ide sequence "TATAC"
corresponds to a reference seguence "TATAC" and is complementary to a reference sequence "GT~TA."
The terms " ubstantially corresponds to", "substantially homologous", or "substantial identity" as u~d WO94/0056g ~ ;-c~ PCT/US93/05~

herein denotes a characteristic of a nucleic acid sequence, wherein a nucleic acid sequence has at least 70 percent se~uence identity as compared to a reference sequence, typically at least 85 percent sequence identity, and preferably at least 95 percent sequence identity as compared to a reference sequence. The percentage of sequence identity is calculated excluding small deletions or additions which total less than 25 percent of the reference sequence. The reference sequence may be a subset of a larger sequence, such as a portion of a gene or flanking sequence, or a repetitive portion of a chromo~ome. However, the reference sequenGe is at least 18 nucleotides long, typically at least 30 `:
.nucleotides long, and preferably at least 50 to 100 nucleotides long. "Substantially complementary" as used herein refers to a sequence that is complementary to a `
sequence that substantially corresponds to a reference sequence.
Specific hybrîdization is defined herein as the formation o~ hybrids between a targeting transgene seguence (e.g., a polynucleotide of the invention which may include :~
substitutions, deletion, and/or additions) and a speci~ic target DNA sequence (eO~., a human APP gene sequence or human imm~noglobulin gene sequence), wherein a labeled targeting trans~ene seguence preferentially hybridizes to the targ~t s~ch that, for example, a single band corresponding to a : restriction fragment of a gene can be identified on a Southern : blot of DNA prepared from cells using ~aid labeled targeting transgene sequence as a probe. It is eviden~ that optimal hybridization conditions will vary depending upon the saquence c~mposition and length(s) of the targeting transgene(s) and endogenous taryet(s), and the experimental method selected by the practitioner. Various guidelines may be used to select appropriate hybridization condition~ (~ee, Maniatis et al., - Molec~lar Clonin~ A Laboratory Manual (19893, 2nd Ed., Cold Spring Harbor, N.Y. and Berger and Kimmel, M~thod5 in EnzYmoloov~ Volume 152~ Guide tD Molecular Cloninq Techniques (1987), Academic Press, Inc., San Diego, CA., which are incorporated herein by reference.

W094/00s69 PCT/US93/05873 The ~erm "naturally-occurring" as used herein as applied to an object refers to the fact that an object can be found in nature. For example, a polypeptide or polynucleotide sequence that is present in an organism (including viruses) that can be isolated from a source in nature and which has not been intentionally modified by man in the laboratory is naturally-occurring. ~s used herein, laboratory strains o~
rodents which may have been selectively bred according to classical genetics are considered naturally-occurring animals.
The term t'cognate" a-~ used herein refers to a gene ~ sequence that i5 evolutionarily and functionally related between species. For example but not limitation, in the human ,genome, the human immunoglobulin hea~y chain gene locus is the :~
cognate gene to the mouse immunoglobulin hea~y chain gene locus, since the sequences and structures of these two ~enes ~
indicate that they are highly homologous and both genes encode ~: :
a protein which unctions to bind antigens specifically.
As used herein, the term l'xenogenic" is defined in r lation ~ a recipient mammalian host cell or nonhuman animal and means that an amino acid sequence or polynucleotide sequence is not encoded by or present in, respectively, the naturally-occurring genome of the recipient mammalian host cell or nonhuman animal. Xenogenic DNA sequences are foreign DNA s~.quences; for example, human APP genes or immunoglobulin Z5 genes are xenogenic with respect to murine ES cells, also, for illustration, a human cystic fibrosis~associated CFTR allele is xenogenic with respect to a human cell line that is homozygous for wild-type ~normal) CFTR alleles. Thus, a cloned murine nucleic acid sequence that has been mutated (e.g., by site directed mutagenesis) is xenogenic with respect to the murine genome from which the se~lence was originally derived, if the mutated sequence does ~`IOt naturally occur in the ~urine genome.
As used herein, a "heterologous gene" or "heterologous polynucleotide sequence" is defined in relation to the tranægenic nonhuman organism producing such a gene product. A heterologous polypeptide, also referred t~ as a xenogeneic polypeptide, is defined as a polypeptide having an W094/0~569 ~ 3 1 ~ PCT/VS~3/0587 amino acid sequence or an encoding DNA sequence corresponding to that of a cognate gene found in an organism not consisting of the transgenic nonhuman animal~ Thus, a transgenic mouse harboring a human APP gene can be described as har~oring a heterologous APP gene. A transgenic mouse harboring a human immunoglobulin gene can be described as harboring a heterologous immunoglobulin gene. A transgene containing various gene segments encoding a heterologous protein sequence may be readily identified, e.g. by hybridization or DNA
19 sequencing, as being from a species of organism other than the transgenic animal. For example, expression of human APP amino acid sequence9 may be detected in the transgenic nonhuman , animals of the invention with antibodies specific for human APP epitopes encoded by human AP gene se~ments. A cognate ;
heterologous gene refers to a corresponding gene from another species; thus, if murine APP is the reference, human APP is a cognate heterologous gene (as is porcine, o~ine, or rat APP~
along with AP genes from other species).
~s used herein, the term "t~rgeting construct"
refers t~ a polynucleotide which comprises: (1) at least one homology region having a sequence that is substantially identical to or substantially complementary to a sequence present in a host cell endogenous gene locus, and ( 2 ) a : targeting region which becomes integrated into a host cell endogenous gen:e locus by homologous recombination between a targeting construc~ homology region and said endogenous gene lscus sequence. If the targeting construct is a 'thit-and-run"
or "in-and-outt' typ2 construct (Valancius and Smithies 51991) Mol. Cell. Biol~ 1402; Donehower et al. (1992~ Nature 356:
215; (1991~ J NIH Res. 3: 59; Hasty et al. (1991) Nature 350;
243, which are incorporated herein by reference), the targeting region is only transiently incorporated into the endogenous gene locus and is eliminated from the host genome by selectîon. A targeting region may comprise a seguence that is substantially homologous to an endogenous gen~ sequenc~
and/or may comprise a nonhomologous se~uence, such a~ a s~lectable marker (e.g., neo, tk, gpt). The term "targeting construct" does not necessarily indicate that the ~ 1 3 ~ 3 1 t~
W094/00~69 PCT/US~3/05873 polynucleotide comprises a gene which becomes integrated into the host genome, nor does it necessarily indicate that the polynucleotide comprises a complete structural gene sequence.
As used in the art, the term "targeting construct" is synonymous with the term "targeting transgene" as used herein. -~
The terms "homology region" and "homology clamp" as used herein refer to a segment (i.e., a portion) of a targeting construct having a sequence that substantially ~orresponds to, or is substantially complementary to, a prede~ermined endogenous gene sequence, which can include sequences flanking said gene. A homology region is generally at least about 100 nucleotides long, preferably at least about .25Q to 500 nucleotides long, typically at least about lOOQ
nucleotides long or longer. Although there is no demonstrated theoretical minimum length for a homology clamp to mediate homologous recombination, it is believed that homologous recombi~ation efficiency generally increases with the length of the homology clamp. Similarly, the recombination efficiency increases with the degree of sequence homology between a targeting construct homology region and the endoyenous target sequence, with optimal recombination efficiency occurring when a homology clamp is isogenic with : the endogenous target se~uence. The terms "homology clamp"
: and "homology region" are interchangeable as used here , and the alternative terminology is offered-for clarity, i ew -.^
the inconsistent usage of similar terms in the art.
homology clamp does not necessarily connote formation a base-pair~d hybrid structure with an endogenous sequence.
Endogenous gene sequences that substantially correspond to, or are substan~ially complementary to, a transgene homology region are referred to herein as "crossover target sequences"
or "endogenous target sequences."
As used herein, the term "minigene" or "minilocus"
refers to a heterologous gene construct wherein one or more nonessential segments of a gene are deleted with r~pect to ::
the naturallyToccurring gene. Typica}ly, del~ted segments ar~ -i~tronic sequence~ of at least about 100 basepai~s to several - kilobases, and may span up to several tens of kilobases or W094/00569 , PCT/US93/0587 more. Isolation and manipulation of large (i.e., greater than about 50 kilobases) targeting constructs is frequently difficult and may reduce the efficiency of transferring the targeting construct into a host cell. Thus, it is fre~uently desirable to reduce the size of a targeting onstruct by deleting one or more nonessential portions of the gene.
Typically, intronic sequences that do not encompass essential regulatory elements may be deleted. For example, a human immunoglobulin heavy chain minigene may compxise a deletion of an intro~ic segment between the J gene segments and the ~
constant region exons of the human heavy chain immunoglobulin gene locus. Frequently, if convenient re^ctriction sites bound . a nonessential intronic sequence of a cloned gene sequence, a deletion of the intronic sequence may be produced by: (1) diyesting the cloned DNA with the appropriate restriction en~ymes, (2) separating the restriction fragments (eOg., by electrophoresis~, (3) isolating the restriction fragments encompassing the essential exons and regulatory elements, and (4) ligating the isolated restriction ~ragments to form a minigene wherein the exons are in the same linear order as is present in the germline copy of the naturally-occurrin~ gene.
Alternate methods for producing a minigene will be apparent to those of skill in the art (e.g., ligation of partial genomic clones which encompass essential exons but which lack portions of intronic sequence). Most typically, the gene segments comprising a minigene will be arranged in the same linear order as is present in the germline gene, however~ this will not always be the case. Some desired regulatory elements (e.g., enhancers, silencers) may be relatively position-insensitive, so that the regulatory element will functioncorrectly even if positioned differently in a minigene than in the corresponding ge~mline gene. For example, an enhancer may be located at a different distance from a promoterl in a differ~nt orientation, and/or in a different linear order. For example, an enhancer that is located 3' to a promoter in germline configuration might be located 5' to the promoter in a minigene. Similarly, some genes may have exons which are alternatively spliced at the RNA level, and thus a minigene W094/00s69 PCT/US93/OSB73 may have fewer exons and/or exons in a different linear order than the corresponding germline gene and still encode a functional gene product. A cDNA encoding a gene pro~uct may also be used to construct a minigene. However, since it is generally desirable that the heterologous minigene be expressed similarly to the cognate naturally-occurring nonhuman gene, transcription of a cDNA minigene typically is driven by a linked gPne promoter and enhancer from the naturally-occurring gene.
A~ used herein, the term l'large transgene" or "large homologous targeting construct" generally refers to polynucleotides that are larger than 50 kb, usually larger .than 100 kb, frequently larger than 260 kb, occasionally as large as 500 kb, and sometimes as large as 1000 kb or larger.
As used herein, the term "transcriptional unit" or "transcriptional complex" refers to a polynucleotide s2quence that comprises a structural gene ~exons), a cis-acting linked promoter and other cis-acting sequences necessary for efficient transcription of the structural sequences, distal regulatory elements necessary for appropriate tissue-specific and de~elopmental transcription of the structural sequences, and additional cis se~uences important for efficient transcription and translation (e.g., polyadenylation site, mRNA stability controlling-sequences).
As used herein, "linkedl~ means in polynucleotide linkage (i.e~, phosphodiester linkage). "Unlinked" means not linked to another polynucleotide sequence; hence, two sequences are unlinked if each sequence has a free 51 terminus and a free 3' terminus.
DET~ILED DESCRIPTION OF THE INVENTION
Generally~ the nomenclature used hereafter and the laboratory procedures in cell culture, molecular genetics, and nucl~ic acid chemis-ry and hybridization described below are those well known and commonly employed in the art. Standard techniques are used for recombinant nucleic acid mekhods, polynucieotide synthesis, cell culture, and transgene incorporation (e.g., lipofection protocols). Generally W094J0~569 2 1 ~ ~ 3 1~ PCT/US93/~587 enzymatic reactions, oligonucleotide synthesis, and purification steps are performed according to the manufacturer's specifications. The techniques and procedures are generally pexformed according to conventional methods in :
the art and various general references which are provided throughout this document. The procedures therein are believed to be well known in the art and are provided for the convenience of the reader. All the information contained therein is incorp~rated herein by reference.
Chimeric targeted mice are derived according to Hogan, et al., Manipulatinq the Mouse_Embr~o: A Laboratory Manual, Cold Spring Harbor Laboratory (1988) and . Teratocar~inomas and Embryonic Stem Cells: A Practical Ap~roach, E.J. Robertson, ed., IRL Pr~ss, Washington, D.C., (1987) which are incorporated herein ~y reference.
Embryonic stem cells are manipulated according to published procedures (Teratocarcin~mas and Embryonic 5tem Cells_ ~A Practical Approach, E.J. Robertson, ed., IRL Press, Washington, D.C. (1987); Zjilstra et 2~., Nature 342:435 438 (1989); and Schwartzberg et al., Science 246:79~-803 (1989), each of which is incorporated herein by reference).
Oligonucleotides can be synthesized on an Applied : ~ Bio Systems oligonucleotide synthesizer according to specifications provided by the manufacturer.
25 : It has often been observed that cDNA-based transgenes are poorly expressed or inappropriately regulated.
Genomic DNA-based transgenes (i.e., constructed from cloned : genomic DNA sequences) which substantially retain ~he cQntent and organization of the naturally-occurring ge~e locus are more likely to be correctly expressed, but are limited in size by the cloning capacity of bacteriophage and plasmid/cosmid ~ ;
vectors. The yeast artificial chromosome (YAC) i~ a recently ~ :
developed cloning vehicle with a capacity of approximately 2 megabases (Mb~ (Burke et al. (1987) Science 236: 806). The ability to reproducibly and effici~ntly introduce YACs into transgenic mic~ can significantly surpass current transgen~
size limit~

2135.~
W094/00569 PCT/US93/~5873 In general, the invention is based on the unexpected finding that large (i.e.~ greater than about 50 kb) cloned polynucleotides can be efficiently transferred into mammalian cells, such as ES cells, and are incorporated into at least one chromosomal location and stably replicated as a segment of a chromosome. Further, it was found that large cloned polynucleotides comprising a complete transcriptional unit can be transferred into mammalian cells (e.g., ES cells~, incorporated into a chromosomal location, and transcribed to produce a detectable concentration of RNA transcripts of the struc~ural gene sequences. It was also found that unrearranged immunoglobulin genes cloned in YACs can be ,introduced into ES cells and developed to form a transgenic animal in which productive VDJ rearrangement occurs, and expression of immunoglob~lin chains also occurs. It has also been found that large transgenes can be cloned in YACs and, after isolation from ~ e host yeast cells, efficiently transferred into mammalian cells (e.g., ES cells3 without prior separation of the desired transgene sequences from yeast-derived YAC sequences, and that the presence of such :~ yeast-derived YAC sequences can be non-interfering (i.e., : compatible with efficient transgene integration and : transcription of a transgene transcriptional unit).
Unexpe tedly, it also has been fou~d that large transgenes, with or without linked yeastderived YAC sequences, can be efficiently co-transfected into mammalian cells ~e.g., ES
cells) with unlinked polynucleotides containing a selectable marker, such as, for example, a ~eoR expression cassette; and that selection for cells harboring the selectable marker gene 30 and expressing the selectable marker are are highly likely to also harbor the large transgene species which has been co-lipofected, thus allowing efficient selection for large transgene DNA sequences without requiring prior ligation (and cloning) of a selectable marker gene. The finding that large DNA segments, such as YAC clones, can be efficiently co-lipo~ected with a selectable marker gene permits, for the first time, the construction of transgenic mammalian cells and .transgenic nonhuman animals harboring large xenogenic DNA

wo g~oos~g 2 ~ 5:~ 1 3 PCT/US93/0587~

segments that are typically difficult to manipulate. Thus, large polynucleotides, typically 50 to 100 kb in size, frequently more than 250 kb in size, occasionally more than about 500 kb, and sometimes 1000 kb or larger, may be efficiently introduced into mammalian cells. The mammalian cells may be ES cells, such as murine ES cells (e.g., the AB-l line), so that the resultant transgenic cells can be injected into blastocyst~ to generate tran~genic nonhuman animals, such as transgenic mice or transgenic rats, harboring large DNA
10 transgenes, which are preferably expressed in the nonhuman transyenic animals. The present methods may also be carried out with somatic cells, such as epithelial cells ~e.g., keratinocytes), endothelial cells, hematopoietic cells, and myocytes, for example.
Embryonic Stem Cells If embryonic stem (ES) cells are used as the transgene recipients, it is possible to develop a transgenic animal harboring the targeted gene(s) which comprise the integrated targetlng transgene(s~. Briefly, this technology involves the introduction of a gene, by nonhomologous ~integration or homologous recombination, in a pluripotent cell line (eOg., a murine ES cell line) that is capable of differentiating into germ cell tissue.
: A large transgene can be nonhomoloyously integrated into a chromosomal location of the host genome. Alternatively, a homologous targeti~g construct (which may comprise a transgen~) that contains at least one altered copy of a :~:
~portion of a germline gene or a xenogenic cognate gene (including heterologous genes) can be introduced into the 3 0 genome of embryonic stem cells . In a portion of the cells, the introduced DNA is either nonhomologously integrated into a :
chromosomal location or homologously recQmbines with the endogenous ~i~e., naturally occurring) copy of the mouse g~ne, - replacing it with the altered construct~ Cells containing the 35 newly engineered genetic sequence(s) ar~ jec:~ed into a host mouse blastocyst, which is reimplanted into a recipient female. Some of these embryos develop into chimeric mice that possess a population of germ cells partially derived from the ~094/00569 2 1`3 5:3 1 ~ PCT/U~93/~873 mutant cell line. Therefore, by breeding the chimeric mice it is possible to obtain a new line of mice containing the introduced genetic lesion ~reviewed by Capecchi et al. (1989) Science 244: 1288, incorporated herein by referenc~).
For homologous targeting constructs, targeting efficiency generally increases with the length of the targeting transgene portion (i.e., homology region) that i5 .substantially complementary to a reference se~uence present in the target DNA (i.e., crossover target seguence). In ~eneral~
targeting efficiency is optimized with the use of isogenic DNA
homology regions, although it is recognized that the presence of recombinases in cer~ain ~S cell clones may reduce the .degree of sequence identity required for effici~lt recombination.
The invention also provides transgenes which encode a gene product that is xenogenic (e.g., heterol~gous) to a nonhuman host species. SuGh transgenes typically comprise a structural gene sequencP expression cassette, wherein a linked :promoter and, preferablyr an enhanc~r drive expression of structural sequences encoding a xeno~enic (e.g., heterologous protein). For example, the invention provides transgenes : which comprise a mammalian enhancer and at least one human APP
promoter linked to structural sequences that encode a human -~
: APP:protein. Transgenic mice harboring such transgenes express human APP mRNA(s). Preferably, the polynucleotide sequen~e encoding the xenogenic (e.g., hetPrologous) protein is operably linked to cis-acting transcriptional regulatory regions ie.g., promoter, enhancer) so that a heterologous protein is expressed in a manner similar to the expression of : :
the cognate endogenous gene in the naturally-occurring nonhuman animal 7 Thus, it is generally preferab~e to operably link a ~ransgene structural encoding sequence to transcriptional regulatory elements which naturally occur in ~ .
-- or near the cognate endogenous gene. However, transgenes encoding h~terologous proteins may be targeted by employing a homologous gene targeting construct tar~eted adjacent to the endogenous transcriptional regulatory sequences, so that the operable linkage of a regulatory seq~ence occurs upon W~94/00569 2 1 3 5 8 ~ ~ ~ PCT/US93/05~

integration of the transgene into a targ~ted endogenous chromosomal location of th~ ES cell.
Selectable Marker Genes A selectable marker gene expression cassette typically comprises a promoter which is operational in the targeted host cell ~e.g., ES cell) linked to a structural sequence that encodes a protein or polypeptide that confers a selectable ph~notype on the targeted host cell, and a polyadenylation signal. A promoter included in an Pxpression cassette may be constitutive, cell type-specific, stage-specific, and/or modulatable (e.g., by hormones such as glucocorticoids, MMTV promoter), but is expressed prior to and/or during selection. An expression cassette can optionally include one or more enhancers, typically linked upstream of the promoter and within about 3-10 kilobases.
However, when the selectable marker is contained in a homologou targe~ing ccnstruct, homologous rec~mbination at the targeted endogenous site(s) can be chosan to place the selectable marker structural se~uence downstream of a functional endogenous promoter, and it may be pcs~ible for the targeting construct replacement region to comprise only a ;:;
structural sequence encoding the sele~table marker, and rely ::
upon an endogenous promoter to drive transcriptio~ (Doetschman et al. (1988) Proc. Natl Acad. Sci. (U.S.A.l 85: 8583,~:
incorporated hsrein by referencP). Similarly, an endogenous enhancer located near a targeted endogenous site may be relied on to enhance transcription of selecta~le marker gene sequences in enhancerless constructs. Preferred expre5sion cassektes of the invention encode and Pxpress a selectable drug resistance marker andJor a HSV thymidine kinase enzyme.
Suitable dru~ resistance genes include, for example: gpt ~xanthine-guanine phosphoribosyltransferar-e), which can be selected for with mycophenolic acid; neo (neomycin phosphotransferase), which can be selected for with G418, hygromycin, or puromycin; and DFHR (dihydrofolate reducta~e) t which can be selected for with methotrexate (~ulligan and Berg (1981) Pro~. Na~l. Acad. Sci. (U.S.A.~ 78: 2072; Southern and Berg (1982) J. Mol. Ap~l. Genet. 1: 327; which are 213~3~ 3 ` ~:
W094/0056~ PCT/US93/05~73 incorporated herein by ref~rence). Other suitable selectable markers will be apparent to those in the art.
Selection for correctly co-lipofected recombinants will generally employ at least positive selection, wherein a selectable marker gene expression cassette encodes and expresses a functional protein (e.g., neo or gpt) that confers a selectable phenotype to targeted cells harboring the endo~enously integrated expression ca~sette, so that, by addition of a selection agent (e.~ 418, puromycin, or mycophenolic acid~ such targeted cells have a growth or ~ survival advantage over cells which do not have an integrated expression cassette. Further guidance regarding selectable .marker genes is available in several publications, including Smith and Berg ~1984) Cold Sprinq Harbor Sym~. Ouant. Biol.
49: 171; Sedivy and Sharp (19~9) Proc. Natl. Acad. Sci.
(U.S.A.3 86: 227; Thomas and Capecchi (1987) op.cit., which are incorporated herein by reference.
Larqe Xeno~enic Polynucleotides Large polynucleotides are usually cloned in YAC
vectors. For example, human genomic DNA libraries in YAC
cloning vectors can be screened (e.g., by PCR or labeled polynucleotide prob~ hybridization) to isolate YAC clones spanning complete genes of interest ~e.g., a human APP gene, a human immunoglobulin heavy cha~n locus or light chain locus)~
25 ~ or ~ignificant portion~ of such genes which comprise a complete transcriptional unit. Methods f~r making YAC
libraries, isolating desired YAC clones, and purifying YAC DNA
are described in the art (U.S. Pa~ent 4~889,806; Burke et al.
(1987) Science 236- 806; Murry et al. (1986) Cell 45: 529, incorporated herein by reference~
Once a desired YAC clone is isolated, and preferably deproteinized, yeast-derived YAC sequences may optionally be completely or partially removed by digestion with one or more restriction enzymes which cut outside the desired cloned large trans~ene sequence; yeast-derived sequences are separated from thP clo~ed insert se~uences by, for example, pulsed gel electrophoresis. Preferably, a complete unrearranged YAC

W094/005~9 2 1 3 ~ 3 1 ~ PCT/US93/05 clone is used as a large transgene or large homologous targeting construct in the methods of the invention.
In one aspect, preferred YAC clones are those which completely or partially span structural gene sequences selected from the group consisting of: human APP gene, human immunoglobulin heavy chain locus, human immunoglobulin light chain locus, human al-antitrypsin gene, human Duchenne muscular dystrophy gene, human Huntington's chorea-associ2ted loci, and other large structural genes, preferably human g~nes.
Preferred YAC cloning vectors are: a modified pYAC3 vector ~Burke et al. (19~7) o~.cit., incorpora~ed herein by . . .
.referen~e), pYACneo (Traver et al. (1989) Proc. Natl. Acad.
ci. (U.S.A.) 86 5898, incorporated herein by reference), and pCGS966 (Smith et al. (1990) Proc. Natl. Acad._Sci. (U.S.A.
87: 8242, incorporated herein by reference).
Cationic Li~ids Lipofection, and various variations of its basic methodology, have been described previously in th~ art ~U.S.
Patents 5,049,386; 4,946,787; and 4,897,355) and lipofection reagents are now sold comme~cially (e.g., "Trans~ectam9' and ~:~
; "Lipofectin"). ~`~
Cationic and neutral lipids that are suitable for ::
; efficient lipofection of DNA have been described in the art.
~ipofection may be accomplished by forming lipid complexes with DNA made according to Felgner (W091/17424, incorporated : herein by reference~ and~or cationic lipidization ~W091/16024; ~.
incorporated herein by reference). Various lipofection protocols described in the art may be adapted for co- ~-lipofection according.to the invention; for example but not ~;
limitation, general lipofectlon protocols ar~ described in ~he fo~lowing references which are incorporated herein: Behr et al. (1989) Proc. Natl. Acad. Sci. Lu.s.A~! 86: 6982; Demeneix et 1. (1991) nt. J. Dev. Biol. 35: 481; Loe~fler et al, . ~1990~ J. Neurvchem. 54; 1~12; Bennett et alO (1992) Mol.
~21~ 41: 1023; Bertling et al. (1991) ~otechnol. Appl.
Biochem. 13: 390; Felgner et al. (1987) Proe. Natl. Acad. Sei.
(U.S.A.) 84: 7413; Felgner and Ringold (1989) Nature 337: 387;

W094/00~69 2 i 3 5 3 1 3 PCT/US93/05873 Gareis et al. (1991) Cell. Mol. Biol. 37: 191; Jarnagin et al.
(1992) Nucleic Acids Res. 20: 4205; Jiao et al. (1992) Exp.
Neurol. 115: 400; Lim et al. (1991) Circulation 83: 2007;
Malone et al. (1989) Proc. Natl. Acad. Sci. ~U.S.A.~ 86: 6077;
Powell et al. (1992) Eur. J. Vasc. Sura. 6: 130; Strauss and Jaenisch (19~2) EMBO J. 11: 417; and Leventis and Silvius (1990) Biochim. Biophvs. Acta 1023: 124.
Newer polycationic lipospermines compounds exhibit broa~d c:ell ranges ~Behr et al., (1989) o~.cit. ) and DNA is coated by these compounds. In addition, a combination o~
neutral and cationic lipid has been shown to be highly efficient at transfection of animal cells and showed a broad .spectrum of effectiveness in a variety of cell lines (Rose et al., (1991) BioTechni~ues 10:520) 15- A lipofection complex (or a cationic lipidized DNA
complex) i5 defined as the product made by mixing a suitable cationic lipid composition with one or more polynucleotide peci~s, such as a large transgene and a selectable marker gene expression cassette. Such a co-lipofection complex is characterized by an înteraction between the polynucleotides and lipid components that results in the formation of a co-lipofection complex that, when contaoted with mammalian cells under suitable conditions (e.g., buffered saline or ES cell medium with or without serum, 20-45C), results in incorporation of the polynucleotides into the ma~malian cells;
preferably thP mam~alian cells are ES cells, such as murine ES
cells.
Various suitable cationic lipids may be used, either a}one or in combination with one or more other cationic lipid species or neutral lipid species. Generally, sl~itabl~
cationic lipids comprise a positively charged head group (one or more charges) and a covalently linked fatty acid tail. A
suitab}e cationic lipid composition is "Trans~ectam" ~ProMega, ~adison, WI) compri5ing the cationic lipid-polyamine dioctad~cylamidoglycyl spermidine (DOGS). DOTMA is a preferred lipid known as N-(2,3-di(9-~3~-octadecenyloxy))-prop-1-N,N,N trimethylammonium chloride. DNA--DOTMA complexes made essentially from DOTMA and DNA. Other axamples of wo g4/00569 2 1 3 5 3 i 3 PCT/US93/0587~

suitable cationic lipids are: dioleoylphosphatidylethanolamine (PtdEtn, DOPE), dioctadecylamidoglycyl, N-trimethylammonium chloride, N-trimethylammonium methylsulfate, DORI and DORI-ether (DORIE). DORI is N-[1-(2,3-dioleoyl)propyl]-N,N-dimethyl-N-hydroxyethylammonium acetate and DORIE is N-[l-(2,3~dioleyloxy~propyl]-N,N-dimethyl-N-hydroxyethylammonium acetate. DOTAP is N- r 1- ( 2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl sulfate; this lipid has ester rather than ether linkages and can be metabolized by cells.
lQ Optionally, one or more co-lipids may be combined ~ with a suitable cationic lipid. An optional co-lipid is to be understood as a structure capable of producing a stable D~A-lipid complex, alone with DNA, or in combination with other lipid components and DNA, and is preferably neutral, although :
it can alternatively be positively or negatively charged.
Examples of optional co-lipids are phospholipid-related materials~ such as lecithin, phosphatidylethanolamine, lysolecithin, lysophosphatidylethanolamine, phosphatidylserine, phosphatidylinositol, sphingomyelin, cephalin, cardiolipin, phosphatidic acid, cerebrosides, dicetylphosphate, dioleoylphosphatidylcholine ~DOPC), dipalmitoyl-pho~phatidylcholine (~PPC), dioleoylphosphatidylglycerol (DOPG), dipalmitoylphosphatidylglycerol (DPPG), diolsoyl- :
phosphatidylethanolamine (DOPE), palmitoyloleoy- ~ :
lphosphatidylcholine (POPC),palmitoyloleoyl~
phosphatidylethanolamine (POPE) and dioleoyl-phosphatidylethanolamine 4-(N-maleimidomethyl) ~ :~
cyclohexane-~-carboxylate ~DOPE-mal). ~dditional non-phosphorous containing lipids are, e.g.,stearylamine, dodecylamine, hexadecylamine, acetyl palmitate, glycerolricinoleate, hexadecyl stereate, isopropyl myristate, amphoteric acrylic polymers, triethanolamine-lauryl sulfate, alkyl-aryl sulfate polyethyloxylated fatty acid amides, dioctadecyldimethyl ammonium bromide and the like.
G~nerally, to form a lipofection complex, the polynucleo ide(s) i5/are comblned according to the teaching~
in the art and herein with a suitable cationic lipid, in the 2i35~13 W094/00569 PCT/US93/~Sg73 presence or absence of one or more co-lipids, at about pH 7.4-7.8 and 20-30C. When more than one species of polynucleotid are combined in a lipofection complex it may be referred to herein as a co-lipofection complex. A co-lipofection complex yenerally compri es a large polynucleotide ta transgene or homologously targeting construct) and a selectable marker gene expression cassette. The co-lipofection complex is administered t~ a cell culture, preferably murine ES cells, under lipofection conditions as described in the art ~nd herein.
General Methods A preferred method of the invention is to transfer a .sub~tantially intact YAC clone comprising a large heterologous transgene into a pluripotent stem cell line which can be used to generate transgenic nonhuman animals following injection ~ into a host blastocyst. A particularly preferred embodiment ~-o~ the inventi~n is a human APP gene targeting constrUct co-lipofected with an unlinked positive (e.g., neo~ selection expression cassette. The human APP transgene is transferred into mouse ES cells (e.g., by co-lipQfection with nQo) under ~onditions suitable for the continued viability of the co-: lipofected ES cells. The lipofected ES cells are culturedunder selective conditions for positive selection (e.g., a selective concentration of G418). Selected cells are then ::~
verified as having the correctly targeted transgene recombination by PCR analysis according to standard PCR or . :
Sou hern blotting methods known in the art (U.S. Patent 4,683,202; Erlich et al., (1991) Science 252: 1643, which are ~:
incorporated herein by referenre~.
Correctly targeted ES cells are then transferred into suitable blastocyst hosts for generation of chimeric transgenic animals according to methods known in the art (Capecchi, M. (1989) TIG 5:70; Capecchi, M. (19B9) Sc~ienc~e 244:1288, incorporated herein by reference). Several studies have already us d ~CR to successfully identify the desired transfected cell lines ~Zimmer and Gruss (1989) Nature 33~:
150; Mouellic et a~. (1990) Proc. Natl. Acad. Sci. (U.S.A~l 87: 4712; Shesely et al. (1991~ Proc. Natl. Acad. Sci. US~ ~8:

W~94~00569 2 1 3 5 3 1 3 PC~/USg3/05~--429~, which are incorporated herein by reference). This approach is very effective when the number of cells receiving exogenous ~argeting transgene(s) is high (i.e., with electroporation or lipofection) and the treated cell populations are allowed to expand (Capecchi, M. (1989) -op.cit., incorporated herein by referencP).
For making transgenic non-human animals (which includ~ homologously targeted non-human animals), embryonal stem cells (ES cells) are preferred. Murine ES cells, such lQ as AB-l line grown on mitotically inactivP SNL76/7 cell feeder .
layers (Mc~ahon and Bradley, Cell 62:1073-1085 (1990)) essentially as desrribed (Robertson, E.J. (1987) in Teratocarcinomas and Embryonic ~tem Cells: A Practical ~ :
Approach. E.J. Robertson, ed. (Oxford: IRL Press), p. 71-112) may be used for homologous gene targeting. Other suitable ES
lines in~lude, but are not limited to, the E14 line (Hooper et alO ~19~7) Nature 325: 292-295), the D3 line (Doetschman et al. (1985) J. ~mbryol. Exp. Mor~h. 87: 27-453, and the CCE
line ~Roberts~n et al. (1986) Nature 323: 445-4481. Rat, hamster, bo~ine, and porcine ES cell lines are al~o a~ailable in the art for producing non-murine transgenic non-human animals bearing a human APP gene sequence. The success of generating a mouse line from E5 cells bearing a large transgene or specifically targeted genetic alteration depends on the pluripotence of the ES cells (i.e., their ability, once injected into a host blastocyst, to participate in embryogenesis and contribute to the germ cells of the resulting animal). The blastocysts containing the in~ected ES
cells are allowed to develop in the uteri of pseudopregnant nonhuman females ~nd are ~orn as chimeric mice. The resultant transgenic mice are chimeric for cells having the large transgene(s~/homologous targeting constructs and are backcrossed and screened for the pr~senc. of the transgene~s) andlor YAC sequences by PCR or Southern blot analysis on tail biopsy DNR of offspring ~o as to identify transgenic mice heterozygous for the transgene(s)/homo~ogous targeting constructs. By performing th~ appropriate cros~es, it i5 possible to pro~uce a transgenic nonh~man animal homozygous 213531~
~W094/00~69 . - PCT/US93/0~873 for multiple large transgenes/homolo~ous recombination constructs, and optionally also for a transgene encoding a different he~erologous protein. Such transgenic animals are satisfactory experimental models for various diseases linked to the transferred transgene~s).
- For performing certain types of studies, trans~enic rats harboring and expressing a human APP sequence may be preferred. :
EXPERIMENTAL_EXAMP~ES .-~

Materials and Transfection Calibration :
Pilot experiments to determine toxicity levels, optimum DNA:lipid ratios, etc. were performed with ] (2kb PGKneo cassette in pUC) and with pYPNN (a modified pYACneo vector containing a PGKneo cassette in place of the SV40-neo ~
cassette in the acentric arm. ~.
: The YAC used in these calibration experiments was an 85kb human IgH gene fragment cloned into a modified pYACneo vector (EcoRI >NotI loning site alteration). The YAC was thus lOOkb in length including the vector arms.
: DOTMA (Lipofectin, BRL, Bethesda, MD) and DO~S
(Transfectam, ProMega, Madison, WI) were tested as cationic ipids. Fig. 1 shows chemical structures of representative cationic lipids which can be used to form co-lipofection ~co~ lexes. ES cell toxicity curves were performed for each.
: : Toxic effects could be seen with DOTM~ at the 30~g/ml level~
DOGS sho~ed no toxic effect-~ at the 60~g/ml level.
: Optimal DNA:lipid ratios were determined for both lipids, using pGKneo as reporter. Optima for DOTM~ and DOGS
were at 1:10 and 1:50 (DNA:lipid, wt:wt~, respectively. :
Optimal ES cell number for ~ach lipid was determined. Optimal incubation times and conditions were .
determined (how long and in what buffer gave maximal trans~ection with ~oth DNA:lipid complexes). Pilot studies indi~ated that 3-5 x 106 ES cells incuba~ed with ~ 1:50 mix~ure o~ DNA:DOGS in serum free DMEM for 3-4 ~ours at 37C
yielded the maximal number of transfectants. T~ese conditions were routinely used for the YAC lipofertion experiments.

W~94/00569 2 1 3 ~ ~ 1 3 PCT/US93~0~

Neor was pro~ided by an unlinked plasmid carrying a PGXneo cassette, either eem or pYPNN, in a co-lipofection procedure. DNA:plasmid molar ratios varied from 1:4 to 1:20.
An equal wPight of carrier DNA (sheared herring sperm) was also added.
Yeast blocks were prepared at 3.5 x 109 cellstml in 0.67% low gel temp agarose. The YAC was isolated by PFGE.
Outer lan2s were stained with EtBr and aligned with the unstained portion. A thin slice containing the YAC was :
isolated using a brain knife. Approximately l~g of YAC was recovered in ~pproximately 10mls of gel. The gel was washed extensively in g~lase buffer (40mM bis-Tris pH 6.0, lmM EDTA, .40mM NaCl), melted at 70C, cooled tG 40~C, and incubated with 10U g~lase (Epicentre Technologies, Madison, WI) overnight.
After digestion, pYPNN or picenter were added at a typical molar ratio of 1:4 (YAC:plasmid)~ An equal weight of sheared herring sperm DNA was added as carrier. TAe agarase dig~stion mix ~ontaining approximately 100 mg of the YAC wa~
directly incubated with Transfectam at optimal DNA:lipid ratios for 30 minutes at room temperatures. No polyamines were added.
ES cells were washed, trypsinized, and resuspended in serum free DMEM. Nine mls of cell suspension containing 3 x 106 ES cells (and about 105 feeder cells) were plated onto a 60mm petri plate tnot tissue culture plastic~. About 1 ml of the DNA-lipid mix was added to the cells in DMEM and incubated at 37C for 3-4 hours. The cells were then collected in DMEM
FBS, and plated at 106 per 100mm tissue culture dish. G418 selection was applied 24 hours later, and colonies picked after about 10 days.
Typically, 1-2~g of YAC was used per experiment.
Thus, about 10-20 separate lipofections were performed on a given day. Generally, several hundred G418 resistant clones - were picked~ of which at lPast 1-2% contained speci~ic YAC
derived sequence. Of these, approximately 10% carry the intact YAC, as determined by fine structure south~rn blotting using probes covering the entir~ YAC ins~rt, PF~E southern analysis, and PCR analysis.

W094/0~569 2 1 ~ 5 3;1 3; P~T/US93/05873 Production of mice carryin~_a YAC encodinq human AmYloid Precursor Protein Preparation of the APP YAC DNA: A 650kb human genomic fragment containing the full length APP gene was isolated as a yeast artificial chromosome (YAC) in a yeast host strain (clone #B142F9) from the Washington University YAC
library (available from Center for Genetics in Medicine Librarian, Washington University School of Medicine, St.
Louis, Missouri). The yeast strain was grown to late log phase in AHC medium, resuspended in 0.67% low gelling e~perature agarose (SeaPlaque, FMC Corp.) at 3.5 x 109 .cells/ml, and cooled in block formers (Bio~Rad). Intact yeast chromosomal DNA was prepared as follows. Thirty 250~1 blocks of B142F9 ~ells were swirled in a 150 mm petri dish containing 50ml~ of YSS ~YSS: 4mg/ml NoYozyme 234 (Novo Nordisk), lM
sorbitol, lOOm~ EDTA, 50mM Potassium Phosphate, pH 5.5) at :37C for 30 minutes. The blocks were wa~hed once in TE (10 mM
Tris pH 7.5, lmM EDTA) and swirled in 50 mls o~ LDS ~LDS: 1%
lithium dodecyl sulfate, 1~ sarcosyl, 100 mM E~TA~ for 30 to 60 minutes at 37C. The LDS was removed with a sterile 50ml pipette, and the blocks swirled in 50mls of fresh LDS
overnight. The blocks were rin ed several times in 50mM EDTA, and stored at 4~C in 50mM EDTA. 100~1 segments of the prepared blocks were loaded into each well of a 1~ low gelling temp agarose gel in 0.25X TBE 114 x 25cm CHEF gel, 10 well gel comb, Bio-Rad). The yeast chromosomes were separated by pulsed ~ield gel electrophoresis (CHEF-DRIL, Bio-Rad) using a 60 second switch time at 200V and 14C for 4~ hours~ The end lanes of the gel were removed, stained for 2 hours in 0.5~g~ml ethidium bromide, and the separated chromosomes visualized on a W transluminator. Under these condition-~, the 650kb YAC
was separated from the nearest endogenous yeast chromosome ~y - 3-5mm. The gel segments were notched to indicate the location of 650kb YAC, and the segments realigned with the remainder of the gel. A 2 mm wide slice of the gel containing the 650kb YAC was isol~ted using a brain knife ~Roboz Surgical Instrument Co) and stored in 50mM EDTA at 4 C. Approximately W094/~0~69 2 1 3 5 3 1 ~ PCT/US93/05~ ~

5 ~g of YAC DNA was isolated in approximately 10 mls of gel.
The agarose slice containing the YAC DNA was equilibrated in gelase buffer (40mM bis-Tris pH 6.0, lmM EDTA, 40mM NaCl), melted at 70C for 20 minutes until completely liquid, and cooled to 45C. Gelase (25 U, Epicentre Technologies, Madison, WI) was added and the molten agarose mixture was incubated at 45~C for 90 minutes to liquify the D~A-agarose mixture.
Introduction of the APP YAC into ES cells Embryonic stem cells ~AB-l) were washed in PBS, trypsinized, and resuspended in serum free ES medium (DMEM, lX
glutamine, pen/strep, 1 mM 2-mercaptoethanol, lX NEAA).
.Approximately 5 x ~o6 cells in 9 ml of serum-free ES cell suspension were placed into each of ten 60mm petri (non-tissue culture treated) dishes. A linearized plasmid containing a selectable marker (PGKneoA+R, containing the PGK promoter fused to the neomycin resistance gene; R~dnicki et al. (1988) Mol. Cell. Biol. 8: 406; Rudnicki et al. (1989) Biochem~ Cell.
Biol. 67: 593, incorporated herein by reference) was added to ~0 1 ml of gelase treated YAC DNA at a 2:1 (plasmid:YAC~ molar ratio. A cationic lipid (Transfectam, ProMega, Madison, WI) was added a~ a 50:1 (Transfectam:DNA) weight:weight ratio, the mixture was gently inverted once to mix and incubated ~t room temperature for approximately 30 minutes. One ml of the D~A:lipid mixture was then added to each 60mm dish of ES cells and incubated for 4 hours in a 37C CO2 incu~ator. The cells were then transferred to a sterile 250ml bo~le, an equal volume of ES medium ~as above, but including 15% fPtal calf serum3 was added. Cells were removed from the dishes with gentle pipetting and combined with an e~ual volume of ES
medium containing 15 percent fetal calf serum~ This cell suspension was transferred in 15 ml aliquots to 100mm tissue culture plates containing mitotically inactivated SNL75/7 fibroblast feeder cells (McMahon and Bradley (1990~ Cell ~:
1073, incorporated h~rein by re~erence~ and returned to the ki~sue culture incubator for 24 hours~ After 24 hours, the medi~m was changed to ES medium con aining 10 percent fetal calf serum and 400~y/ml G418, and refed every 48 hours. After ~1~53~
.WO9~/0~569 - PCT/US93/0~873 7 days, a t~tal of 366 G418 resistant colonies were counted.
Each of 240 colonies were individually transferred to a well of a 96-well microtitre dish containing 50 ~1 of 0.25 percent :
trypsin in calcium-free magnesium-free PBS. After 15 minutes, 50 ~1 ~f serum-containing medium was added, the colony dissociated by trituration, and the cell suspension was transferred to duplirate 96-well plates containing culture medium and feeder layers (as supra) . A~ter 4~ 5 days, one set of dishes was frozen according to conventional methods ~Ramixez-Solis et al. Guide to Techniques in Mouse DevPlop~ent (199~) Methods in EnzYmsloav, incorporated herein by reference). Cells were dissocia~ed in 50 ~1 trypsin and mixed . with 50 ~1 of Freezing Medium (20% DMSO, 20% fetal calf serum in DMEM3; 100. ~1 of sterile silicon oil was layered on top of the cell suspension in each well, and the plates were placed in Styrofoam containers and frozen at -~0C.
Identification of ES clones containina APP sequences The other set of microtitre dishes containing lipofectant ES clones was used to prepare DNA for PCR
analysis. ~0~1 of lysis buffer (50mM Tris pH ~.0, 200mM ~aCl, 25mM EDTA, 0.2% SDS, lmg/ml Prot~inase K) was added to each well rRamire2-Solis et al. (1992~ Anal. Biochem. 201: 331, i~corporated herein by reference). After an overnight incubation at 55~C, S~l of 2.5M NaCl and 95 ~1 of 100% ~tOH
were added to each well. The dishes were gently swirled at room temperature for 60 minutes to precipitate the D~A. The wells were then rinsed 5 times with 70~ EtOH, and dried at 37C. The DNAs were resuspended overnight in 100~1 of H2O at 37C in a humidified incubator.
The individual DNA samples were pooled in rows and colu~ns for PCR (Fig. 2), and the pools analyzed for APP
sequences by PCR, using the following primers ~adapted from Fidani et al., Human Molecular Genetics 1, 165-16~, 1992):

APP-PA: 5'-GCT TTT GAC GTT GGG GGT TA-3^
. APP-PB2: 5'-TTC GTG AAC AGT GGG AGG GA-3' APP-17A: 5l-ATA ACC TCA TCC AAA TGT CCC C-3 ~PP-17B: 5-GTA ACC CAA GCA TCA T&G AAG C-3 wo g4/00569 2-1 3 5,~ 1 3 P~T/US93/OSY-~4 APP-PA/PB2 denote primers specific for the promoter region of the human APP gene, APP7A/7B are specific for exon 7, and APP17A/17B are specific for exon 17.
PCR analysis of the pools indicated 42 clones which :
potentially carried both promoter and exon 17 sequences (Fig. 3). PCR analysis of the 42 clones individually indica~ed 6 clones (#s 23, 24, 176, 213, 219, 230) containing both promoter and exon 17 sequences (Fig. 4). These clones were expanded in culture, and frozen in vials in liquid nitrogen. The~e cells were mounted in agarose blocks for PFGE
analysis, ~nd harvested for RNA isolation.
Structural~Analysis of E5 clones containinq APP sequences The integrity of th~ APP YAC carried by ES clones was first estimated using a rare cutter fingerprint technique as follows. Restriction enzymes which infrequently cut human DNA were used to define patterns of fragments which hybridized to a human alu fragment probe. Rare cutters often contain the dinucleotide Cp~, and mammalian cells often methylate Cp&
- dinucleotides rendering most restriction sites containing them refractory to digestion. However, yeast cells do not methylate CpÇs, and thus the pattern of CpG containing restriction sites in a given fragment will depend on whet~er the fragment is propagated as a YAC in yeast or within a mammalian cell line. Thus, only rare cutter enzymes without CpG in ~heir recognition sequence were used to generate a diagnostic pattern of alu-containing fragments from the YAC.
Bl42F9 agarose blocks were digested completely with the res~riction enzymes Sfi I, Pac I, Swa I, Pme I, and ~pa I, and analyzed by PFGE Southern blotting using total human DNA
as a probe for Alu fragments (Fig~ 5~. The pattern of bands generatefi by Sfi I digestion was used as a reference pattern, since there was an even distribution of bands from 30 kb to 220 kb. If a YAC were to integrate intact into ES cells, Sfi I digestion would be expected to generate a similar pattern of bands~ with the exception of the terminal fragm~nt~. The terminal fragm~nts could be easily identified ~y reprobing the Sfi I digest with pBR322 sequences. In this case, the entire set of fragments were ordered by par~ial digest mapping.

W094/0~569 2 1 3 5 3 1 ~ PCT/US93~05873 Briefly, B142F9 blocks were digested with a range of Sfi I
concentrations, separated by PFGE, and probed with either the 2.5 kb (trp ~rm specific) or the 1.6 kb (ura arm specific) Bam HI-Pvu II fragment of pBR322, such that at a particular level of partial digestion, a ladder of bands were generated. Each band differed from its nearest neighbor by the distance to the neighboring Sfi I sites (Fig. 6).
The six ES lines were digested to completion with Sfi I and probed under high stringen~y conditions with total human ~NA. Of the six lines, only thre~ (24, 176, 230) showed a pattern of bands consistent with the reference pattern from the APP Y~C. The rare cutter fingerprinting approach does not . require any knowledge of the se~uence of the fragment cloned in the YAC, and is thus applicable to the analysis of any YAC
containing human DNA. Further, if probes for repetitive elements from other species which were not found in the target mammalian cell line were available, this approach could be used to analyze the structure of YACs containing other foreign DNAs into other mammalian cell lines.
Transcriptional analysis of APP YAC containinq ES clones The six ES lines were analyzed for transcription by PCR. Total RNA was prepared from Pach cell line by standard guanidium isothiocyanate/lithium chloride procedures (Sambrook et al., Molecular Cloning). Complementary DNA was prepared using an oligo-dT primer, and the cDNAs were analyzed by PCR
for splice products of the human APP gene using the PCR
primers deri~ed from exons 6 and 9 as depicted below (adapted from Golde et al., Neuron 4 253-267, 1990). To exclude inappropriate amplification of mouse APP cDNAs, the 3' ~nd nucle~tide of each oligo was choseFI such that it was specific for the human cDNA se~uence and not the corresponding mouse cDNA sequence. PCR oligos specific for mouse APP cDNA were also prepared.
human APP sp~cific oligos:
APP-HASl: 5'-CAG GAA TTC CAC CAC AGA GTC TGT GGA A-3' ~PP-HAS2: 5' CAG GAT CCG TGT CTC GAG ATA CTT GTC A~3' , 2~35313 W094~00569 - PCT/US93/05P-' mouse APP specific oligos:
APP-MASl: 5'~CAG GAA TTC CAC CAC TGA GTC CGT GGA G-3' APP-MAS2: 5'-CAG GAT CCG TGT CTC CAG GTA CTT GTC G-3' Clones 24, 176, and 230 showed the expected PCR
bands indicative of alternatively spliced human APP
transcripts encoding the 770, 751, and 695 amino acid forms of the protein (Fig. 7). Clones 23, 213, and 219 did not contain PCR detectable txanscript, and also serve~ as a negative control, indicating that the human APP specific oligos did not amplify bands from mouse APP transcripts endogenous to the ES
cell lines. Further, the RT-PCR analysis confirmed and validated the resulks of the rare cutter fingerprint analysis .which predicted that clones 24, 176, and 230 contained the intact YAC whereas clones 23, 213, and 219 did not.
~uantitative anal~sis of APP transcript in ES cells Since RT-PCR analysis is qualitative, RNase protection assays are used to quantitate the alternatively spliced human APP transcripts in the ES lines. The RNase probe was generated by cloning the 310 bp RT-PCR product as an E~o RI-Bam HI fragment into the vector pSP72 (Fig. 8). The resultant plasmid~ pHAPP, is linearized at the Hpa I site, and ~-:
: :antisense transcript is generated from the SP6 promo er.
RNase protection assays are performed according to standard protocols (Sambrook et al., Molecular Cl~ning~.
~lterna~ively, Sl nuclease protection analysis is : used to quantitate the transcripts. pHAPP is digested with Xho I and Hpa I to release a 446 bp fragment. The double stranded fragme~t is end-labelled with Klenow, denatured, hybridized to ~NA samples from the ES lines carrying human APP
sequences, and Sl analysis performed according to standard m~tho~s (Sambrook, et al., Molecular Cloning~.
In addition, expression of human APP protein can be determined by immunoprecipitation of human APP using antibodies specific for human APP protein from ES cell lin~s and tis~ue of tran5qenic animals. Such ~ntibodies may also permit direct detection o~ human APP by standard immunohistochemical analysis of tissue sections Anal~sis of human APP ex~ression in transqenic mice 2135~

The qualitative and quantitative assays described above are also applicable to the analysis of the human APP
gene in tissues of the transgenic mice derived from these ES
lines.
Production of chimeric founders and qermline transmission of he_APP YAC
Clones 23, ~13 and 219 were injected into blastocysts to generate chimeric founder animals as described (~obertson, ed. Teratocarcinomas and Embryonic Stem Cells).
1~ Founders ara bred to wild type mice to generate F1 animals carrying the ~PP YAC.
Mou e models of Alzheimer's Disease ~verexpression of the wild type human (or mouse) APP
protein may result in phenotypes characteristic of Alzheimer's Dlsease, including neurofibrillary tangle formativn, plaque formation, and neurological dysfunction. Toward that end, di~ferent mouse lines expressing the~APP YAC can be interbred to increase the number, and hence expression, of human APP
g nes.
Alternatively, mutations identified as associated with Familial Alæheimer's Disease (codon 717:V->I, F, or G) and/or Hereditary Cerebral Hemorrhage with Amyloidosis, Dutch type ~HCHWA-D, codon 693:glu->gln) may be introduced into the human APP gene contained on the YAC using standard yeast molecular genetic techniques such as in~ertion/eviction of a plasmid carrying a subcloned fragment of the APP gene containin~ the mutation, or by oligonucleotide directed transformation of yeast (Guthrie and Fink~ Guide to Yeast Molecular Genetics and Molecular Biology). YACs ~.arrying : .
these mutated APP genes can be introduced into transgenic mice using procedures described a~ove. O~her naturally-occurring human APP disease allele sequences also may be used, including but not limited to those described at codon 692 (Ala -> Gly) t - codon 692 ~Glu -> Gln or Gly), and codon 713 (Ala ~> Val) and `~
others that are described (Hendricks et al. (1992) Nature Genet. 1: 218; Jones ~t al. (1992) ature Genet. 1: 306; Hardy et alO ~1992) ~ature Genet. 1: 233, Mullan et a}. (1992) W~94/00569 2 1 3 5 3 1 ~ PCT/~Sg3/05~ -Nature Genet. 1: 505; Levy et al. (1990) Science 248 1124, incorporated herein by reference).

Transqenic Mice ExPressinq a Human Immunoqlobulin Gene Cloned in a Yeast Artificial Chromosome An 85 kb fra~ment of the human heavy chain immunoglobulin ge~e was cloned as a YAC, and embryonic stem cell lines carrying substantially intact, integrated YACs were derived by colipofection of the YAC and an unlinked selectable marker. Chimeric founder animals were produced by blastocyst injection and offspring transgenic for the YAC
~lone were obtained. Analysis of serum from these offspring for the presence of human heavy chain demonstrated expression of the YAC borne i~munoglobulin gene fragment. Unlike fusion of yeast spheroplasts with mammalian cells, no yeast chromosomal DN~ need be introduced by the co-lipofection method a6 the YAC(s) are typically first isolated ~rom yeast chromosomes by a separation method, such as pulsed field gel electrophoresis (PFGE). The YAC was introdu~ed into ES cells by co-lipofection with an unlinked selectable marker plasmid.
The co-lipofection strategy differs from lipofection of modified YACs in that retrofitting vectors do not need to be constructed or recombined into the YAC, and YACs carried in recombination deficient hosts can be used. In contrast to microinjection approaches, it is likely that larger YACs can be introduced by co-lipofection than microinjection due to the technical hurdles in purification o~ intact YAC DNA and because of the high shear forces imparted on the DNA during microinjection. Furthermore, unlîke fusion of yeast : 30 spheroplasts with mammalian cells where some of the yeast ;
chromosomes integrate with the YAC5~ 6, no yeast chromosomal DNA is introduced in c~-lipofection since the YAC is first isolated by pulsed field gel electrophoresis.
Transgenic mice were produced by blastocyst injection of ~S cells carrying an intact YAC. The YAC was maintained intact through the germline, and human heavy chain antibody subunits were detected in the serum of transgenic offspring.

W094/0~s69 ~l.J J ~, ~ PCT/USg3/05~73 Human Heavy Chain Gene Fragment The 85 kb Spe I fragment of the unrearranged human immunoglobulin heavy chain locus was isolated. The 85kb Spe I fragm~nt of the human heavy chain immunoglobulin (H) chain gene contains at least one of each element required for correct rearrangement and expression of a human IgM heavy chain molecule.

An 85kb Spe I restriction fragment of the human hea~y chain immunoglobulin gene contains VH6, the functional diversity tD) segments, all six joining (J) segments, and the C~ constant region segment ~Hofker et al. (1989) Proc natl.
. Acad. ~Sci. uy~S.A.) 86: 5587; ~erman et al. ~1988) EMBO J~ 7:
727; Shin et al. ~1991) EMBO J. 10: 3641). Fresh human sperm was harvested and genomic DNA prepared in agarose blocks as described in Strauss et al. (1992) Mamm. Ge.nome 2: lSO). A
size selec~ed (50-lOOkb) Spe I complete dig~st YAC library was ~ :
prepared in the yeast host strain AB1380 in pYACneol5, u~ing the Spe I site near the entromerP as the cloning siteO A
: 20: ~ize selected (50-lOOkb) Spe I complete digest YAC library was produced in the YAC Yector pYACneo15 and screened by colony hybridization with a probe specific for human C~ ~Traver et al. ~1989) Proc. Natl. Acad. Sci. ~U.S.A.~ 86: 589B). one positive cl~ne (Jl~ was identified among approximately 18,000 pri~ary transformants. Because yeast mitochondrial DNA often obscured the YAC on pulsed field gel electrophoresis, a r-petite variant làcking mitochondrial DNA was selected by EtBr `
: treatment, and denoted Jl.3P. One subclone, Jl.3P, was mounted in agarose~blocks at 3.5 x 109 cells/ml and intact yeast chromosomal DNA was prepared (5mith et alO (1990) Proc.
Natl. Acad. Sci._ ru.s~A~ 8242). The YAC DNA was isolated in a 3-4mm wide gel slice from a low melting point preparative C~EF gel (Biorad). The gel slice was equilibrated in b-agarase buffer (Gelase, Epicentr~ Technologies), melted at 70-C for 20 minutes, cooled to 45-C, and digested with lO
units o f agarase overnight at 45 C.
Characterization _ ~ YAC Jl.3P

WO9~/00569 2 1 ~ 5 3 ~ ~ PCT/USg3/0~ -The authenticity of the J1.3P insert was determined by restriction mapping and Southern analysis. The ends of the insert were subcloned, using the bacterial selectable markers in the centromeric and acentromeric arms of pYACneo. Fine structure restriction analyses of the terminal fragments were entirely consistent with published maps and sequences of the region (Fox et al. nalysis and maniPulation of yeast mitochondr-al qenes, In Guide to Yeast Genetlcs and Molecular Biology (1991) eds. Guthrie C and Fink G, Academic Press, San Diego~ California; Word et al. (1989) In~._Immunol. 1: 296) and defined the orientation of the insert with respect to the vector arms. The orientation was further verified by PCR
: analysis of the acentromeric insert for VH6 sequences, and hybridization of the centromeric insert with the C~ probe.
Southern analysis of the C~ region was consistent with published maps and restriction analyses ~Hofker et al. (1989) Proc._Natl. Acad. Sci._(U.S.A.L 86: 5587). The functional diversity segments of the human heavy chain are contained in a 35 kb ~pan containing a four-fold polymorphic repeat of D
segments. Southern analysis of the J1.3P YAC produced a "restriction fraqment fingerprintl' of the D region in which al~ of the D specific bands in the YAC were present in human genomic DNA.
~: Co-lipofection of Jl.3P YAC into ES cells -: 25 The J1.3P YAC was co-lipofected with an unlinked linearized plasmid carrying the neor gene driven by the mouse PGK promoter (Soriano et al. (l991~ Cell 64: 893)o Selectable marker lasmids Plasmid is a 5 kb plasmid containing an expression casse~te consisting o~ the neo gene under the transcriptional control of the mouse phosphoglycerate kinase-l promoter and the PGK-l poly ~A) site ~Tybulewicz et al. (1991) Cell 40~
271). The plasmid pYPNN is a variant of pYACneo containing - the PGKneo cassette in place of the SV40 promoter-neor cassett~, conætrurted by exchange of a 4.5kb Sal I-Apa I
fra~ment of pYACneo for a 1.5kb Sal I-Apa I fragment of a containing the PGK promotor, neor coding region, and the W094~00569 2 1 3 5 ~ 1 ~ PCT/US93~05873 PGXp(A) signal. The plasmids were linearized with Sal I ~a~
or Not I (pYPNN).
Lipofection of YAC DNA into ES cells.
The digested agarose/DNA mixture was divided into 1 ml (approximately 100 ng) portions in polystyrene tubes (Falcon) and 100 ng pYPNN or 20 ng , and 1 ~g sheared herring sperm DNA (Sigma) was mixed in each tube, and cationic lipid tTransfectam~ ProMega) was then added at a 10:1 ratio (wt:wt~ and gently mixed into the DNA solution. The mixture 10 wa~ incubated for 30 min at room temperature to allow formation of DNA-lipid complexes. Rapidly growing conf luent cultures of AB-1 embryonic stem (ES) cells on mitotically ~;
. inacti~ated SNL 76/7 fibroblast feeder layers were trypsinized ~;
to yield a single cell suspension, washed with serum~
containing medium, and resuspended in serum-fr2e DMEM (Gibco).
For each lipofection~ 9 ml of cell suspension containiny 3 x 1O6 ES cells and about 1 x 105 feeder cells w~re mixed with 1 ml of the DNA-lipid m~xture in a 60 mm petri dish (Falcon 1007; Becton Dickinson) and incubated for 4 hours at 37 C in 20 ~ ~a humidified 5% CO2 atmosph~re. Dishes were swirled gently .
during the incubation to minimize cell attachment. After incubation, cells were diluted with serum-containing ES cell medium, dispersed gently, and plated at 1 x 106 on 100 mm culture dishes containing feeder layers. Cells were selected ~: :
~5 in G418 (400 ~g/ml powd2r, Gibc~) for 9-lZ days, beginning 24 ~ ~ :
: hours after plating. Two different plasmids were tested: pYPNN
(a 12 kb derivative of pYACneo carrying the PGKneo cassette in place of the SV40-neo cassette) and ickensian (a 5 kb plasmid carrying the same PGKneo cassette). The YAC:plasmid molar 3~ ratio was 1:8 for pYPNN and 1:4 for ickensian. Two cationic lipid formulations were tested, DOGS (Transfe::tam; ProMega~
and DOT~ (Lipo~ectin; BRL). Similar trans~ection efficiencies were obtained for DOGS and DOTMA wit~ linearized plasmids, but DOGS was ultimately chosen for the YAC
experiment5 because its cationic mQiety is spermine, obvi~king the need for exogenously added spermine as a DNA protectant, ~nd because DQGS waS not toxic to E5 cells at the concentrations used. Because the DNA:lipid ratio was found to W0~4/00569 ~ 1 35~1~ PCT/US93/05~--be important to the transfection efficiency, and precise measurement of the YAC DNA concentration was difficult, each lipof~ction contained an estimated 10-fold excess ~1 ~g) of sheared herring sperm carrier DNA to provide a baseline level of DNA.
Analysis of ES clones G41~-resistant clones were dispersed with trypsin an~ the cells from each clone were dividPd into one well of a 96-well plate that was frozen and a second 96-well or 24-well plate used for preparation of DNA for screening by Southern analysis. Positi~e clones were thawed and expanded for further analy~is.
Southern blot hybridization and PCR
Genomic DNA was prepared from ES cells and tail biopsies by ra~id preparation methods ~Laird et al. (1991) Nucleic Acids Res. 19: 4293) and subjected to South~rn analysis by standard methods. For pulsed field gel electrophoresis, ES cells were embedded in agarose blocks at 1O7 cells/ml, prepared for restriction digPstion, and dige~ted overnight with Spe I. For Southern analysis of pulsed field gels, the DNA was acid-nicked, then transferred to GeneScreen Plus ~DuPont) in denaturing solution ~0.4N NaOH, 1.5 M NaCl).
51igonucleotides suitable for PCR amp.ification of the VH6 region were prepared from published sequences. Primers used were 5'CAGGTACAGCTGCAGCAGTCA3' and 5'~CCGGAGTCACAGAGTTCAGC3', which amplified a diagnostic 275 bp product.
Production and analysis of trans~enic mice Clones containing intact YAC sequences were injected into blastocysts to produce chimeric founder animals, which were bred with C57BL/6 wild ~ype mice and JH- mice, which carry targeted inactivations of ~oth ropies of the mouse heavy chain gene. Thymic cells from transgenic offspring were mounted in agarose blocks for pulsed field gel eleGtrophoresis and Southern analysis to confirm transmission of the intact YAC.
ELISA assays Human mu chain was detected using a 2-site ELISA
assay. Polyvinyl chloride microtiter plates were coated with W094/00569 PCT/US93/~5~73 mouse monoclonal anti-human IgM clone CH6 (The Binding Site, San Diego, CA) at 1.25 ~g/ml in 100 ~l PBS by overnight incubation at 4 C. Plates were blocked by l hr incubation with 5% chicken serum (JRH, Lexana, XS) in PBS. Following 6 washes with PBS, 0.5% tween-20, serum samples and standards were di.luted in 100 ~l PBS, 0.5% Tween-20, 5% chicken serum (PTCS) and incubated in the wells for 1 hr at room temp.
Purified human myeloma-derived IgM, kappa (Calbiochem, La Jolla, CA) was used as a stand rd. Plates were then washed 6 times with PBS, 0.5% tween-20 before addition of peroxidase ~`
~ conjugated rabbit anti-human IgM, Fc5u fragment specific antibody diluted 1/lO00 in lO0 ~l PTCS. After another 1 hr . incubation at room temperature, the wells were washed 6 times and developed for 112 hr with 100 ~l ABTS substrate (Sigma) .
Assay plates were read at 415-490 nm on a Vmax microplate - reader (~olecular Device~, Menlo Park, CA), and IgM
cQncentration determined from a 4-parameter logistic curve fit of the standard values. A level of 4.89 ng/ml in serum samples i5 routinely detected by this assay and differentiated from bac~ground by at least 3 standard deviations.
Results Approximately eight ~g of Jl.3PYAC DNA were lipofected in eight separate experiments (Table 1). Of the 1221 G418 resistant clones screened, 15 contained the two diagnostic Eco RI C~ fra~ments (Table 2). Two of the clones ~#s 195 and 553) contained only one of the two C~ bands, which may have arisen from fragmentation of the YAC within the missing Eco RI fragment (Table 2~. The two selectable marker plasmid~, pYPNN and ace, produced, respectively, frequencies of 0.5 and 13.5 G418 resistant clones per 106 transfected cells; the efficiency of pYPNN selection was much lower eYen though it was used at twice the molarity of ace. It is unclear why the plasmids differed, since they both contained the same neor cassette, but it may be a con~equence of the 35 . extent of se~uence homology ~etween the plasmid and the vector armst or different efficiencies of neor expression from the two plasmids.

WO 94/00569 2 1 3 5 3 1 3 ~ `/US93/05~ `

~ t) o o a)- u~
~z ~n o ,t `~ h.C
~ .,1 0 :3~ ~D ~'X ~ S
0 ~ Inh O
Q) O :
C~ ~ U

I V ~ ~ ~1 0 0 0 0 r~ ~ ~ a'V

~ ~ ~'~
a c ~ C~ o o ~ 0 e~

o ~,1 ,,, o ~
~: 3 . U C:: O
_I a~
t~ ILO
c: aJ~r~ O o O ~n ns O h I ~ , :~ o t) 00 o~
I a~
f~5 ~ ~ 3 1 -~
O O ~::
1 ~ .~
I " ~ ~
Cl: ~ Cl ~ f~
~ ~,C
C~
I
.-~ o O' I ~S U~
a .~ co oo co O ~
I¢ q-l 3 ._.~ O
f. I

I s: Q) O
O ~
O
~) ,4 ~a ~: O O O
~1 ~ ~Z; Z Z; Z o u7 z; æ ~ æ ~ Q, u~ ~4 Q~ ~n ~ ~ ~ ~ ~ ~ ~ ~ ,~ 3 ~: Z
I o a a) ~ o r~ U
c~ ~ ~
,~ ~ o ~1 ~ O
Q; q:

~C .~ o E~ O Q3 ~n :
s 213531:~
WO ~4~00569 PCI`/USg3/05873 ~ ~X
Q, U~ 0~ ~ H
,~ O X ~
'CS O

3~Q ~ U~
C ,C S ~
~a 0 ~ u~ o ~ Q) Q.~ ~ ~n .~ ~ o ~ ~ GS ~ u o-~ ~ H O
~ Ull ~ + + + I ,~ ~ C~: S O ~ ~ 3 ~,~
_~ Q~v ~ ~ ~
Ir)¦ 'C5N C~ N IU O ~U h ,~ ~S
I I I I C ~ Q) C)~ ~I R O 3 `~
~ ~~ Q) O ~O
a 0 Q) ~ ~ C ~l ri .C ~ ,~
O a) ~ ~ ~ ~ ~ o a I + ~ + a~ i r C a~
r~ , C ~ o ~o ~ s ~ o o ~ o-~
:~ ~ 3 ~ ~,~q ~J ~,G~ '8 V Q

~1 1 1 + ~ s o 1: ~al ~
O ~1 ~ ~ +
--~ ~ a ~ s ~ ~ Q
u~ ~ I ~ a) h ~ 0 ~
~C Q) U) ~ Q) ` ~? ~ ~: X h ~ S.J
tn~ ) 01) N
~1 ~11 ~ ~ + + I ~ ~ ~ Q 0~
O ~ ~ 0 10 C ~5 C CQ ~ ~q ~Dl ~ O O
S~4 ~I C 11~ ~
~ ~ + .c a~ ~ + Cl' o Z
::1 ~ ~ U~ ~ 3 ~
U: ~i ~ rl ~ ~ ~ V
tl ~ I + + C U~ A ~ U3 S~ J H U~
o ~
. u~ ~ O
~ ~1 ~ ~ + + ~ o ~: ~

-~ ~1 ~ ~ + ~ o ~ u ~ ~o Pt;
Q 3 :> ~.ÇI .C ~: O U
C ~- X ~ Q) ~ ~- ~
~I t` + ~ ~ t ~.)~rl ,IC: q.l 'O t~S O ~U ~ -I Pb .~ Q ~C
+ ~ ~ a) - (5 ~ ~1 ~ ~ ~ ..
` + U~ ~ N ~
O ~ R
U U~ + ~4 G) ~
~ ~ Q, ~ O O
,~ 5 0 ~ ~q U
Q) ~ Q) ~ J h al ~ h C) ~ rl G ~ * ~
C .~ i ~ ~ 3 a) ~ t '~ O
I E~ ~ /11 0.c ~ o o ~ e ~ o J~
o ~ o ~ o ,l ~ ~

W094/00569 ~ ~ 3~3 1 3 PCT/US93/05~-Analysis of YAC structure in ES cells.
The four C~+ clones from the pYPNN co-lipofections (#s 12,14,18,21) were analyzed for D region structure by the restriction fragment fingerprint assay de~cribed above. Of the four clones, only clone 18 retained the fingerprint of the parent YAC. Clones 14 and 21 contained fewer bands than the parent, suggestiny that YAC sequences may have been lost, ~-while clone 12 contained several additional bands, consistent ~;
with integration of more than one copy of the YAC in this ES
line.
The integrity of the 3 ' end of the insert region in the four ES ~ines wa~ assessed by Southern analysis using the 10.5kb Nde I-Spe I terminal fragment isolated by vector recircularization as probe. Three bands are expected from a Xho I digest of the parent YAC: a verv large D-J-C~ band (>30 kb), a i.5 kb C~-C~ band, and an 8.9 kb C~-vector band. A
double digest with Xho I and Spe I is-expected to reduce the size of the 8.9 kb ~and to 4.1 kb. The 4.5 kb and ~.9 bands are pres~nt in the Xho I digests, while the 4.5 kb and 4.1 kb bands are present in the Xho I-Spe I digests of the parent ; ~Y~C. Among the four ES lines, only line 18 contained the parental YAC banding pattern indicati~e of an intact 3'-end.
The presence of an 8~9 ~b band is consis~ent with the ~retention of the vector arm Xho I site in the ES line, sugges~ing that very little of the telomeric region had been lost in this clone. Loss of YAC terminal sequences would ~e ~: expected to result in aberrant Xho I bands. Among the other three ES lines, clone 14 lacked the 4~5 kb Xho I band, while clones ~2 and 21 contained aberrantly short Xho I bands, indicating rearranged or deleted 3l end regions in these ES
clones. A similar analysis of 5' end integrity was not possible due to repetitive elements in the region. However, PCR and Southern analysis using the Y~6 PCR product a~ probe indicated that clone 18 contained VH~ sequences, while clones 14, 12, and 21 did not (Table 2). ::
O~ the 13 C~+ ES cell lines from the d co- :
lipofectiorls, one was lost during clonal expansion, and one ~ :
266) was eliminated because it lacked VH6 sequence. The -W094/00569 2 1 3 5 3 1 3 PCT/VS93/~873 remainder were analyzed for D region structure, 3'-end integrity and/or VH~ sequence (Table 2). Of the 11 lines analyzed for D region fingerprint, six (#s 86, 191, 220, 371, 463, 567) showed an intact D region while five had aberrant patterns. 3' end analysis of five of the six lines with intact D regions revealed that all but one (~220) contained an intact 3' end. PCR analysis revealed that f iYe of the six line with intact D regions (#86, 220, 371, 463, 567) contained V~6 sequences, while only one of five lines without intact D regions (#35) contained VH6 sequences.
Ten of the ES cell lines were examined for full length insert by pulsed field Southern analysis using tha ~
, r~gion or C~ probe (Table 2). Only clones 18, 371, and 463 contained an 85 kb Spe I fragment indicative of a full length insert; all of the other clones had a smaller Spe I fragment.
The Spe I digest of clone 18 was screened with both D and C~
probes and a probe for VH6; all three probes hybridlzed to a single band of 8S kb.
The pulsed field Southern analysis, taken together 2Q with the D region, 3' end and VH6 fine s~ructure analy~es, indicate that the YAC insert was transferred intact in three ES lines: #18, #371, and #463. A high degree of internal rearrangement, deletion or fragmentation was generally seen in the ES lines carrying disrupted YAC se~uences, ~lthough subtle alterations of structure were also detected (e.g., #567~.
Overall, the fre~uency of intact YAC transfer was low, 1 în 400 G41~ clones (3/1221). However, the isolation of the : clone DNAs and the primary screen for C~ sequences (which eliminated 1206 of the 1221 clones from further analysis) were rapidly performed using the microtitre plate protocols desrribed in Methodology. Thus, only 15 clones required extensive analysis (Table 2).
Molecular analysis of YAC stxucture in ES cells is grea ly ~acilitated by a low, preferably single, copy of ~he ..
YAC. The D region, pulsed field gel analysis, and 3' end ànalyses of the ES ~ines are consistent with a low or single copy integration of the YAC. Analysis of clones 18, 371, and 463 for a diagnostic 3' end flanking band showed that clones .-W094~0569 ~ 1 3 5 3 ~ 3 PCT/US93/OS~-' 18 and 371 carriPd a single copy of the YAC insert, while 463 may have an ~dditional intact or partially intact copy.
Production of chimeras and germline transmission of the YAC
Blastocysts were injected with ES lines 18, 371, and S 463. Chimeric founder animals ranging from 10% to 95% ES cell contribution to coat color were derived from all three lines.
The oldest animal, a 40% chimeric male derived from ES line 18, ~ransmitted the ES cell genotype to 20 of 73 offspring.
Eleven of the 20 agouti of~spring were positive for an intact D region fingerprint, consistent with Mendelian segregation of a hemizygous YAC transgene allele. In addition, pulsed field Southern analysis using the D region probe demonstrated a . single 85 kb Spe I band in transgenic offspring, indicating that the YAC was stably maintained through the germline.
Thus, co-lipofection of YACs into ES cells does not abrogate ES cell totipotency.
Southern analysis of integr~tion sites for the co-lipofected selectable marker indicated integration of 2 to lO
plasmid copies. Because it is possible that the marker plasmids could be a source of mutations if they were to insert at multiple loci, the integration sites of the plasmid were tracked by Southern analysis for plasmid sequences. Sin~e pYPNN and the YAC vector arms lack Eco RI sites and contain pBR322 sequences, each Eco RI band which hybridized to a pBR322 probe represent~ the integration of a separate intact : .
or fra~mented copy of pYPNN or the YAC vector arms. Analysis of ES cell clone 18 DNA revealed eight Eco RI bands ranging in : gize from 5.5 kb to 20 kb, and the offspring of a hemizygous transgenic animal bred with non-transgenic mates were analyzed for se~regation o~ the Eco RI bands. Among 14 offspring, all eight Eco RI bands were detected in tail DNAs of the 9 transgenic pups, and none were de~ected in tail DNAs of the non-transgenic pups~ Thus, all detectable marker plasmids segregated with the YAC, indicating that they had inserted at or near the YAC integration site. Co-integration o~ different :~
DNAs have bee~ observed in transgenic mice produced by microinjectîon of zygo~es, and it is expected that co~
integration of plasmid DNAs would be no more mutagenic for co-wo g4~00s6~ 2 1 3 5 ~ 1 3 PCT~US93/Q5873 lipofection than for zygote microinjection. Presumably, the herring sperm carrier DNA had also co-integrated with the YAC, and may be a source of Eco RI sites in the Southern analysis.
Since co-integrated carrier DNA may potentially adversely affect YAC transgene function, it is frequently preferable to omit carrier DNA. Preliminary experiment~ with a 650 kb YAC
indicate that carrier DNA is not required for efficient lipofection of intact YACs into ES ~ells. This preliminary work also suggests that the size limit of YACs which can be successfully co-lipofected into ES cells is at least 650 kb.
5erum expression_of_human immunoglobulins in trans~enic mice Line 18 transgenic mice were assayed for human mu .chain in the serum by ELISA. Human mu heavy chain was detected in the serum of transgenic offspring (Table 3).
Although the human mu serum levels in the transgenics were clearly within the detectable range, they were very low compared to serum levels of endogenous mouse IgM. The low level of transgene expression is due in part to competition from the endogenous heavy chain gene. The transgene was introduced into a background in which the endogenous heavy chain alleles are inactivated, and in this mouse, the human mu serum levels were ele~ated approximately 10 fold (Table 3).

., :
.

WO 94/00569 2 13 5 3 1 3 PCI/US93/05~--Table 3~ Detection of serum human IqM b~ ELISA

Genotype SexAqe at assav Human IqM
(wk~
YAC 18+ F 3, 9, 20 < 5 ng/ml YAC lB~ M 17 12 . 2 ng/ml YAC 18+ F 10 27 . 0 ng/ml YP~C 18+ F 6, 17 < 5 ng/ml YAC 18~ F 4 5 . 8 ng/ml YAC 18+ F 6 lOo 5 ng/ml YAC 18+ M 6 10. 4 ng/ml YAC l~+lJH M 5, 8 165 ng/ml Wild type F ~ < 5 ng/ml Wild type F 6 < 5 ng/ml Wild type F 6 < 5 ng/ml Wild type M 6 < 5 ng/ml Wild type F 34 < 5 nglml Wild type F 34 < 5 ng/ml Wild type M 3 4 < 5 ny/ml Wild type M 34 ~ 5 ng/ml .

Table 3. Blood samples from transgenic animals and controls were analy~ed by ELISA for human IgM at the ages indicated.
All of the transgenic animals are derived from ~ ~ingle clone 1~ ~ounder chimera, and are hemizygous for the YAC (YAC 18+).
3 0 Five c3f the seYen animals in wild type back~round had cletectable human IgM in their serum. The level of detection of the ELISA was 5 ng human IgM~ml serum. The serum human IgM ~ ~:
level was elevated approximately 10-fold when the YAS~
transgene was bred into a background lacking functional 35 endogenous mouse he~vy chain genes (YAC18+/JH~
. ..

W094/00569 2 1 3 ~ 3 1 3 PCT/US93fO5873 . . ~ , .

FACS Analysis of YA5+lJH- Mice Fluorescence-activated cell-sorting (FACS) analysis was performed on mice positive for the YAC containing the 85kb heavy chain gene fragment and homozygous for a functionally disrupted t"knocked-ou~"~ endogenous murine immunoglobulin heavy chain gene by disruption of the JH region by homologous gene targeting. The mice had a single copy of the YAC
transgene and lacked functional murine heavy chain alleles.
The FACS analysis used antibodies to detect human mu chains, among others, and showed that about 60 cells per 10,000 total peripheral lymphocytes from the mice expressed a human mu chain immunoglobulin, This level is approximately 1-2 percent , of the number of cells that express murine mu chains in a wild type (non-transgenic/non-knockout) mouse spleenO
~ACS detected human mu chain expression in cells obt~ined from the spleen and peritoneal cavity of the YAC~/JH-mice.
Although the foregoing invention has been described in some detail by way of illustration and example, for purposes of clarity of understanding, it will be obvious that : certain changes and modifications may be practiced within the , scope o~ the appended claims.
,:

2~5 : :

.:

Claims (38)

WO 94/00569 PCT/US93/0????

52

1. A method for producing a co-lipofected mammalian cell having incorporated multiple heterologous DNA species, comprising the steps of:
forming a co-lipofection complex comprising a cationic lipid, a first polynucleotide, and an unlinked second polynucleotide comprising a selectable marker gene expression cassette;
contacting a mammalian cell with said co-lipofection complex under conditions whereby said first polynucleotide and said second polynucleotide are introduced into same the cell and are integrated into the genome.

2. A method according to Claim 1, wherein said cationic lipid is selected from the group consisting of: N[1-(2,3-dioleoyloxyl)propyl]-N,N,N trimethylammonium chloride; N[1-2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammonium methylsulfate; N-(2,3-di(9-(Z)-octadecenyloxy))-prop-1-N,N,N-trimethylammonium chloride; dioleoylphosphatidylethanolamine (PtdEtn, DOPE); and dioctadecylamidoglycyl spermidine.

3. A method according to Claim 1, wherein the first polynucleotide is at least 500 kb.

4. A method according to Claim 1, wherein at least one of said selectable marker is a drug resistance gene.

5. A method according to Claim 4, wherein said selectable marker is a gene encoding neomycin resistance.

6. A method according to Claim A, further comprising the step of selecting for cells having said selectable marker.

7. A method according to Claim 1, wherein said mammalian cell is a nonhuman embryonal stem cell.

8. A method according to Claim 7, wherein said nonhuman embryonal stem cell is a mouse ES cell.

9. A method according to Claim 8, wherein said cationic lipid is dioctadecylamidoglycyl spermidine (DOGS), said first polynucleotide contains a human APP gene sequence, and said unlinked second polynucleotide contains a neomycin resistance gene.

10. A method according to Claim 1, wherein said first polynucleotide comprises yeast-derived YAC sequences.

11. A co-lipofection complex, comprising cationic lipid noncovalently bound to a first DNA comprising a YAC clone DNA
and to a second DNA comprising a gene encoding a selectable marker.

12. A co-lipofection complex according to Claim 11, wherein the cationic lipid is selected from the group consisting of:
N[1-(2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammonium chloride; N[1-2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammonium methylsulfate; and dioctadecylamidoglycyl spermidine.

13. A co-lipofection complex according to Claim 11, wherein the YAC DNA comprises a human APP gene sequence or a human immunoglobulin gene sequence.

14. A co-lipofection complex according to Claim 13, wherein said cationic lipid is dioctadecylamidoglycyl spermidine.

15. A co-lipofection complex according to Claim 14, wherein said selectable marker gene is neoR.

16. A co-lipofection complex comprising a polynucleotide of at least 50 kb, an unlinked selectable marker gene expression cassette, and a cationic lipid selected from the group consisting of: N[1-(2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammonium chloride; N[1-2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammonium methylsulfate; and dioctadecylamidoglycyl spermidine.

WO 94/00569 PCT/US93/058??

17. A composition comprising a mammalian cell and a co-lipofection complex of Claim 16.

18. A composition according to Claim 17, wherein the mammalian cell is a nonhuman embryonal stem cell.

19. A composition according to Claim 17, wherein the nonhuman embryonal stem cell is a mouse ES cell.

20. A cotransfected mammalian cell produced by a method of Claim 1.

21. A cotransfected mammalian cell produced by a method of Claim 10.

22. A chimeric transgenic nonhuman animal produced by the method of Claim 1.

23. A transgenic nonhuman animal produced by the method of Claim 9.

24. A transgenic mouse comprising a germline genome comprising a xenogenic polynucleotide sequence of at least 50 kb linked to a yeast-derived YAC sequence.

25. A transgenic mouse according to Claim 24, wherein the xenogenic polynucleotide is at least about 500 kb.

26. A transgenic mouse according to Claim 25, wherein the xenogenic polynucleotide sequence encodes a human APP protein or a human immunoglobulin.

27. A transgenic mouse according to Claim 26, further comprising a neoR gene expression cassette.

28. A co-lipofection complex comprising a YAC containing an unrearranged human immunoglobulin gene comprising the 85 kb Spe I fragment, a selectable marker, and a cationic lipid.

29. A transgenic mouse comprising a YAC comprising an unrearranged human immunoglobulin gene having at least one V
region gene, at least one J region gene, and at least one constant region gene.

30. A transgenic mouse of claim 29, wherein the transgenic mouse expresses a human immunoglobulin chain.

31. A transgenic mouse of claim 29, wherein the YAC comprises a 85 kb Spe I fragment of the human heavy chain locus.

32. A method of introducing a heterologous polynucleotide containing a complete structural gene into a mammalian cell, comprising the steps of:

forming a lipofection complex comprising a cationic lipid and a heterologous polynucleotide comprising a complete structural gene;
contacting a mammalian cell with said lipofection complex under conditions whereby said heterologous polynucleotide is introduced into the mammalian cell and are integrated into the genome.

33. A method of claim 32, wherein said heterologous polynucleotide comprises a complete transcription unit comprising a cis-acting linked promoter which naturally occurs adjacent to said structural gene in germline DNA.

34. A method of claim 32, wherein said heterologous polynucleotide is a yeast artificial chromosome.

35. A method of claim 32, wherein said heterologous polynucleotide is at least 650 kb.

36. A method of claim 32, wherein said mammalian cells are pluripotent mouse embryonic stem cells.

37. A method of claim 36, comrpising the further step of selecting mouse embryonic stem cells which have incorporated the heterologous polynucleotide, introducing said cells into mouse blastocysts to produce transgenic mice harboring a germline copy of the heterologous polynucleotide.

38. A method of claim 37, wherein the lipofection complex comprises a cationic lipid selected from the group consisting of: N[1-(2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammonium chloride; N[1-2,3-dioleoyloxyl)propyl]-N,N,N-t r i m e t h y l a m m o n i u m m e t h y l s u l f a t e ;
N-(2,3-di(9-(Z)-octadecenyloxy))-prop-1-N,N,N-trimethylammonium chloride; dioleoylphosphatidylethanolamine (PtdEtn, DOPE); and dioctadecylamidoglycyl spermidine.