Jriiiriiirl of Molecular and Applied Geri.-zrics
`l:32i'—34l © 1982 Raven Press, New York
`
`Transformation of Mammalian Cells to
`Antibiotic Resistance with a Bacterial Gene Under
`Control of the SV40 Early Region Promoter
`
`P. J. Southern and P. Berg
`
`Dcpiirtirierit of Biochemistry, Stanford University Medical Center, Stmiforri, Criiiforiiiri, U.S.A.
`
` ‘_
`
`Summary: A bacterial gene (neo) conferring resistance to neomycin-
`kanamycin antibiotics has been inserted into SV40 hybrid plasmid
`vectors and introduced into cultured mammalian cells by DNA trans-
`fection. Whereas normal cells are killed by the antibiotic G418, those
`that acquire and express neo continue to grow in the presence of G4_l8.
`In the course of the selection, neo DNA becomes associated with high
`molecular weight cellular DNA and is retained even when ce|ls.are
`grown in the absence of G418 for extended periods. Since neo provides
`a marker for dominant selections, cell transformation to 6418 resis-
`tance is an efficient means for cotransformation of nonselected genes.
`Key Words: Antibiotic resistance—Cell
`transformation—DNA
`transfection—Recombinant DNA—Bacterial 361165-
`j—
`
`There are two principal approaches available
`for the introduction of exogenous DNA into
`mammalian cells. Simian virus 40 (SV40) can
`be used as a transducing vector because it can
`replicate vegetatively in primate cells (1) or be-
`come integrated into host chromosomal DNA
`in a wide variety of cells (2). The experimental
`strategy has been to replace various regions of
`the viral genome with cloned segments of DNA
`and to propagate the recombinants with the aid
`of helper viruses in cultured animal cells
`(3-10). Alternatively, exogenous DNA may be
`introduced directly into recipient cells by either
`the calcium phosphate precipitation technique
`
`
`
`(11), DEAE-dextran (12), or microinjection
`(13,14). The pioneering experiments for this
`approach relied on the transformation of TI('
`mammalian cell lines to a TK+ phenotype
`with the herpes simplex thymidine kinase gene
`(15-1?). Subsequently, unrelated DNA se-
`quences have been integrated with the thy-
`midine kinase gene in either linked or eo-
`transformation experiments (13,19). Transfor-
`mation of appropriate mutant cell lines has also
`been demonstrated with cellular DNA se-
`
`quences and this approach has allowed the
`isolation of the chicken thymidine kinase gene
`(20) and the hamster adenine phosphoribosyl
`transferase gene (21). Morphologic transfor-
`mation of normal cells forms the basis of cur-
`
`Received January 4, 1982; accepted February 18, 1982.
`Address correspondence and reprint requests to Dr.
`P.J. Southern, Department of lmmunopathology,
`ggggps Clinic and Research Foundation, La Jolla, CA
`
`rent attempts to isolate cellular oncogenes
`(22,23).
`Unfortunately, experiments that rely upon
`complementation of cell mutations by trans-
`
`327
`
`Merck Ex. 1005, pg 88
`
`Merck Ex. 1005, pg 88
`
`

`
`328
`
`P. J. SOUTHERN AND P. BERG
`
`Brenner)thattheacquisitionandexpressionof
`
`(
`
`the phosphotransferase gene by mammalian
`cells might confer resistance to G418 toxicity.
`Accordingly,
`the phosphotransferase gene
`_
`from Tn5 (designated here neo) (29,3l) was in-
`1
`troduced into the mammalian transcription unit
`of the pSV plasmid vectors. In this arrange '
`ment the SV40 early promoter is 5'-proximal.
`and an intron and polyadenylation signal are
`3'-proximal, to the neo gene (8_,25). Transfec-
`tion of a wide variety of mammalian cell lines
`with these pSV-neo recombinants yields Stable
`transformants that are resistant to G418 at a
`frequency of one transformant per l0*—l0‘
`‘,
`transfected cells. Our data suggest that trans-
`formation results from the acquisition, Inainte~ I
`nance, and continued expression of the neo =
`gene in the cellular genome. Concurrent with
`our experiments, Colbere-Garapin et al.
`(32)
`achieved mammalian cell transformation to
`G418 resistance with recombinant DNA con-
`taining the Tn5 neo gene linked to the herpes
`thymidine kinase promoter DNA segment.
`
`r
`
`I
`
`MATERIALS AND METHODS
`Cells
`
`!
`
`l
`
`A wide range of cultured mammalian cells
`are sensitive to the antibiotic G418 (Table I) l
`and several of these have been used for trans-
`
`formation experiments. The cells were rou-
`tinely maintained in Dulbecco-modified Eagle I
`medium containing 10% newborn calf serum.
`,-
`penicillin and streptomycin, and the indicated
`concentrations of G418 were added to H16
`medium.
`
`Enzymes
`Restriction enzymes were purchased frorlll
`New England Biolabs and Bethesda Research
`Laboratories and digestions were performed
`;
`according to the supplier’s specifications. Tl
`polynucleotide kinase was purchased front‘?-
`New England Nuclear and S1 nuclease from I
`Boehringer Mannheim. T4 DNA ligase andE- ;
`coli DNA polymerase I were kindly provided .
`by S. Scherer, Stanford University.
`
`? J
`
`Merck Ex. 1005, pg 8i?
`
`duced genes are limited by the availability of
`mutant mammalian cell types to serve as gene
`recipients. Dominant-acting genetic markers,
`for example, those that- produce a selectable
`change in the phenotype of normal cells, offer a
`solution to this difficulty. The isolation of
`methotrexate-resistant transformants after
`transfection of normal cells with DNA from
`
`drug-resistant cells (24) exemplifies this ap-
`proach. However, transformation for metho-
`trexate—resistant dihydrofolate reductase is in-
`frequent and, therefore, the utility for cotrans-
`formation with other genes appears limited.
`Recently, our laboratory devised a family of
`SV40 hybrid plasmid vectors to facilitate
`studies of gene transfer and gene expression in
`mammalian cells (8,2S). Plasmids containing
`DNA segments coding for rabbit B-globin (25),
`mouse dihydrofolate reductase (10), and E. colt’
`xanthine-guanine phosphoribosyl transferase
`(8) can induce the synthesis of the corre-
`sponding gene products in mammalian cells.
`The expression of the bacterial gene, gpt, per-
`mits the utilization of xanthine as a substrate
`for the purine salvage pathway and a selection
`can be established in which gpt functions as a
`dominant marker for cell transformation (26).
`In this paper, we describe a second bacterial
`gene which, when incorporated into the same
`family of plasmid vectors, also" provides a
`dominant selective marker for transformation
`
`of cultured mammalian cells.
`
`The selection for transformation of mamma-
`
`lian cells relies on cell killing by an amino-
`glycoside antibiotic, G418 (27). The structure of
`G418 resembles gentamicin, neomycin, and
`kanamycin (28) but, unlike these related com-
`pounds, G418 interferes with the function of
`80S ribosomes and blocks protein synthesis in
`eukaryotic cells (27). These aminoglycoside
`antibiotics can be inactivated by the bacterial
`phosphotransferases, APH(3’)II and APH(3’)I
`encoded by transposons Tn5 and Tn601, re-
`spectively (29). Jiminez and Davies (30)
`showed that yeast could be genetically trans-
`formed to G418 resistance by the phospho-
`transferase gene contained in Tn601 DNA.
`Thus, it seemed feasible (suggested by Sydney
`
`J. Mol. Appl. Genet, Vol. I. No. 4, I982
`
`Merck Ex. 1005, pg 89
`
`

`
`TRANSFORMATION T0 ANTIBIOTIC RESISTANCE
`
`329
`
`Antibiotic G418
`
`Samples of antibiotic G418 were generously
`provided by Dr. P. J. L. Daniels of Schering
`Corporation. Stock solutions containing 4
`mg/ml G418 in 100 mM N-2-hydroxyethylpi-
`perazine-N’-2’-ethanesulphonic acid buffer, pH
`‘L3, were stored at —20°C and added in ap-
`propriate amounts to the cell culture medium.
`The G418 concentration refers to the actual
`
`amount of drug in the solution and takes into
`account that the solid material was only 40-
`50% G418.
`
`DNA Transfection and Selection of
`
`Transformed Cells
`
`Supercoiled plasmid DNA, without added
`carrier DNA, was introduced into tissue cul-
`ture cells (10 pg for approximately 5 X 10“
`cells) using the calcium phosphate precipitation
`technique (11) with the addition of a glycerol
`shock after 4 h (33). About 48 h after exposure
`to DNA, the cells were trypsinized and re-
`plated at a 1:20 dilution. Within 12- 16 h, G418
`was added to the medium at ‘a concentration of
`
`400 peg/ml. The medium plus drug was changed
`every 4 to 5 days. Colonies were first detected
`after about '3’ days in the selective medium and,
`7-13 days later,
`independent colonies were
`trypsinized in cloning cylinders and transferred
`to microtiter wells. When the colonies were
`
`small, the transplanted cells were grown non-
`selectively for an initial 2-3 days. Once estab-
`lished, the clones were expanded to stable cell
`lines in medium containing 400 pg/ml G418. In
`some instances, the initial selection and sub-
`cloning used 400 pg/ml of G418 but the trans-
`formed cells were subsequently maintained in
`200 pg/ml of G418.
`The selection strategy of permitting cell
`growth prior to the addition of G418 was
`adopted because a significant reduction in the
`transformation frequency occurred if G418 was
`added before 48 h. The transfected cells were
`
`replated at lower cell density because G418 is
`most effective against dividing cells. Con-
`sequently, if cells become stably transformed
`
`early after transfection, cell division prior to
`selection may result in overestimation of the
`transformation frequency. Nevertheless, none
`of the transformants from randomly selected
`colonies appeared to have the same organiza-
`tion of the integrated pSV-nee DNA.
`
`Analysis of Transformed Cell DNAS
`for pSV-neo Sequences
`
`High molecular weight cellular DNA was
`extracted as described by Wigler el al. (17),
`incubated with an excess of restriction en-
`
`zyme, and the digests were separated by elec-
`trophoresis in 0.8% agarose gels. After a mild
`depurination reaction (34),
`the DNA was
`transferred from the gel
`to diazobcnzyloxy-
`methyl paper (DBM paper) (35), hybridized
`with radioactively labeled DNA probes (36),
`and radioautographed using Kodak XR5 film
`and Cronex lightning fast intensification screens
`at -70°C (37).
`
`Analysis of Cytoplasmic RNA
`Extracted from pSV2-neo
`Transformed Cells
`
`Cytoplasmic RNA was extracted from
`semiconfluent cultures of transformed cells as
`
`described previously (38). The RNA was sepa-
`rated from contaminating DNA by pelleting
`through cesium chloride and then the poly A*
`RNA fractions were characterized using the
`Weaver-Weissmann variation (39) of the
`Berk-Sharp procedure (40). DNA hybrization
`probes (shown with individual experiments)
`were prepared by labeling appropriate restric-
`tion fragments at their 5’-ends with ['y-‘“'P]-
`adenosine triphosphate and polynucleotide
`kinase (41). The DNA probes were hybridized
`with RNA samples under conditions of DNA
`excess, RNA-DNA hybrids were digested with
`S1 nuclease, and the protected fragments were
`analyzed by gel electrophoresis (39,40).
`
`Protein Labeling and
`Irnmunoprecipitation Reactions
`
`Semiconfluent plates of pSV2-neo trans-
`formed cell lines were labeled for 14 h at 37°C
`
`1. Mat‘. Appl. Genet., Vol. 1, No. 4. I982
`
`Merck Ex. 1005, pg 90
`
`Merck Ex. 1005, pg 90
`
`

`
`l 1
`
`I l
`
`'
`|
`i
`
`
`
`._..1-»_.__.....-
`
`i I1
`
`CV I
`CV1-P
`TC‘!
`COS
`
`LNSV
`HeLa
`K-562
`
`L
`Ltk"
`3T3
`3T6
`PCC4
`F9
`MEL
`
`" So far, no cell line has been found that is naturally
`resistant to G418.
`
`have been tested (Table 1) are killed by G418
`but CV1 and HeLa are unusual because, at
`
`high cell density, these cells may require I0-I4 ,
`days in G418 (400 ,u.g/ml) before the cell killing ’
`can be observed.
`
`Construction of Recombinant
`
`Plasmids Containing neo
`
`‘
`The bacterial transpo son Tn5 encodes a gene
`(neo) whose protein product——a phosphotrans— [
`ferase (APH(3’)II)—confers resistance to the
`;
`kanamycin-neomycin group of antibiotics (31).
`From the studies of the organization of Tn5 i
`DNA, Reznikoff and colleagues (44,45) weft
`-
`able to identify the DNA segment that is essen-
`,
`tial for the expression of neo. The ColEl:'I'n5
`plasmid pRZ112 (44), a deleted form of the or-
`ginal ColEl:Tn5 hybrid plasmid (pRZ102, Fig.
`1), was the source of the neo DNA segment.
`pRZ1 12 DNA was digested to completion with
`Hincll endonuclease and the 2.5 kb neo DNA
`
`.»
`‘
`
`segment was obtained by agarose gel electro-
`phoresis. After ligating a decanucleotide se-
`quence containing the BamHI restriction site
`(Collaborative Research) to the ends of the neo I
`segment (46) the mixture was digested with an
`
`excess of BamHI and Hi'ndIII restriction en-
`
`-
`
`2
`
`donucleases and the resulting 1.4 kb neo DNA
`fragment was purified by gel electrophoresis.
`.
`This fragment, containing HindIII and Baml-ll
`cohesive ends at the 5'- and 3’-ends, respee '
`tively, was inserted between the Hindlll and -
`Baml-II restriction sites in pBR322 DNA. The _|
`resulting plasmid, pBR-neo (Fig. 1), confersre _-
`sistance in E. colt‘ to both ampiciliin anti
`
`E
`
`Merck Ex. 1005, pg 9
`
`with [“H]leucine (200 .uCi{plate, specific activ-
`ity 55 Ci/mmol, New England Nuclear Labo-
`ratory). The soluble proteins were extracted
`from approximately 2 X 10" cells (7) and im-
`munoprecipitated with an antiphosphotransfer-
`asc APi~l(3’)ll serum (provided by J. Davies,
`Geneva). Esclm-iclu'a coli cells (HBl0l) con-
`taining plasmids were grown to approximately
`2 X 10“ cellslml in M9 minimal medium plus
`glucose with supplements of leucine, proline,
`threonine, and thiamine. Samples of the cultures
`(0.5 ml) were washed and resuspended in the
`same medium lacking leucine and then [“H]leu-
`cine was added (200 ttcifml) for 60 min at 37°C.
`Excess unlabeled leucine was added and, after
`washing in M9 medium, the cells were disrupted
`by sonieation. Cell debris was removed by cen-
`trifugation at 14,000 g for 10 min and the super-
`natant was used directly for immunoprecipitation
`reactions. After incubation overnight at 0°C, the
`immune complexes were adsorbed to inacti-
`vated S. mrreu.-r cells (IgGsorb, Enzyme Cen-
`ter. Boston) and removed by centrifugation
`(42). The S. rmreus cells were washed exten-
`sively and the bound proteins were eluted and
`electrophoresed in SDS polyacrylamide gels
`(43). After electrophoresis,
`the gels were
`treated with EN“HANCE (New England Nu-
`clear Laboratory), dried and autoradiographed
`as described above.
`
`RESULTS
`
`Mammalian Cells Are Sensitive to G418
`
`The sensitivity of various cultured cell lines
`to G418 was assessed by plating cells at low
`cell density in microtiter wells in a medium
`supplemented with various concentrations of
`G418. Even at the highest drug concentration
`tested (800 p.g!ml), sensitive cells divided once
`or twice before cytotoxicity was observed. The
`response time for cell killing appears to corre-
`late with growth rate, since the most rapidly
`growing cells are killed in the shortest inter-
`vals. At
`lower concentrations of G418 (100
`pgfml) there is a significant delay but the cells
`are killed eventually. All of the cell lines that
`
`J’. Mot‘. Appl. Genet, Vol‘. I, No. 4, I982
`
`330
`
`P. J. SOUTHERN AND P. BERG
`
`
`
`TABLE 1. G4i8-sensitive mammalian celllilies“
`
`
`
`
`
` Mouse Monkey Human
`
`Merck Ex. 1005, pg 91
`
`

`
`TRANSFORMATION TO ANTIBIOTIC RESISTANCE
`
`331
`
`Sail
`
`HindI I I
`sgm
`
`Hfndlll
`Bglll
`
`Hpal
`
`Hpal
`
`fl-linclll Hfndlll
`Hpal
`By.-‘ll Hincll
`I
`
`I
`
`}
`: ATG
`I-1- neo —I-|
`
`|lRZ112
`
`$ Hfncll CLEAVAGE
`
`Hfndlll
`
`Hirtcll
`
`|Bgl'll
`
`Hincll
`BamHI LINKERS
`Barr.IHI + Hirtdlll CLEAVAGE
`ISOLATION OF nee GENE FRAGMENT
`
`Hfndlll
`lflglll
`
`Baml-ll
`
`INSEFITION mro pBFl322
`SELECTION eon AmpR,iveo"
`
`
`
`FIG. 1. Organization of transposon Tn5 and scheme for the
`construction of recombinant plasmid pElH-neo. The plasmid
`pFlZ1ll2 contains a complete copy of Tn5 inserted into CoIE1.
`The inverted repeat sequences at the ends of the transposon
`are shown as thick lines. The region essential for expression
`olneomycin resistance in E. coil‘ men) is indicated together
`with the ATG codon that initiates the coding sequence of the
`phosphotransferase. Flelevant restriction endonuclease rec-
`ognition sites are included in the diagram. pFlZ112 was de-
`rived from pFlZ102 by partial digestion with Hincll endonu-
`clease and ligation to eliminate a large segment ofTn5 DNA to
`the 3' side of neo (HinclI endonuclease cleaves at Sell and
`Hpal endonucleasfi recognition sites.) Details of the ma-
`nipulations involved with the construction of pBFl-neo are
`given in the text.
`
`neomycin; since the neo DNA segment inter-
`rupts the tetracycline resistance gene, cells
`carrying this plasmid are sensitive to this anti-
`biotic.
`.
`
`The cloned neo DNA segment was readily
`introduced into the plasmid vector-—pSV2 (8)
`(Fig. 2) by excising the B-globin cDNA segment
`from pSV2-BG (25) with HindIII and Bglll en-
`
`donuclease digestion and substituting the neo
`fragment via the corresponding cohesive ends.
`The pSV3-neo and pSV5-neo (Fig. 2) deriva-
`tives were constructed from pS\/2-neo as pre-
`viously described (8,2S). Each of the pSV-neo
`plasmids replicates efficiently in E. roll‘ strain
`HBl0l and confers resistance to ampicillin and
`neomycin. Cloned isolates of each of the re-
`combinant plasmids were shown to have the
`anticipated structures by appropriate restric-
`tion enzyme analyses (data not shown).
`
`Cell Transformation with Recombinant
`
`Plasmids Containing neo
`
`Transfections of Ltk“ cells with pRZll2 or
`pBR-neo plasmid DNA yielded occasional
`G418-resistant colonies {a frequency of about
`one transformant in S x I0“ transfectcd cells).
`
`However, with 3T6 cells as recipient no trans-
`formants have been recovered with pRZll2
`DNA and only one transformant has been iso-
`lated after transfection with pBR-neu (fre-
`quency about one transformant
`in I07 trans-
`fected cells).
`In control experiments, cells
`transfected with either pSV2-gpt (8), pSV2-[3G
`(25) DNA, or mock-transfected without DNA
`have never yielded G418-resistant colonies
`(frequency less than 1
`transformant in 10’
`transfected cells).
`In contrast to the low frequencies of G418-
`resistant transformation with pRZl I2 (ColEl-
`neo) or pBR-neo plasmid DNAS, several dif-
`ferent mammalian cell lines were transformed
`
`to G418 resistance at relatively high frequency
`with the pSV-neo plasmid derivatives (about
`one transformant in 10“ to 10“ transfected
`
`cells). Although the transformation frequencies
`with the different pSV—neo recombinants fall
`within a relatively narrow range, there is a con-
`sistent small difference which reflects the
`
`plasmid’s potential for replication in different
`host cells (Table 2). For example,
`the fre-
`quency of stable G418-resistant monkey cell
`transformants is two- to three-fold lower with
`
`pSV3-neo than with pSVS-neo, whereas the
`converse applies to the formation of G418-
`resistant mouse cell transformants. This differ-
`
`}. Moi‘. Appl. G(’l'l'£'l‘.. Vol. l'. No. 4, I982
`
`Merck Ex. 1005, pg 92
`
`Merck Ex. 1005, pg 92
`
`

`
`332
`
`P. J. SOUTHERN AND P. BERG
`
`Amp“
`
`
`o_ri
`
`Amp“
`
`pBR322 cri
`
`tr
`Pu II
`SV40 on
`
`pBR322 ori
`
`vu
`[I
`P
`SV40 ori K. H.-ndlll
`
`0.1‘.-'
`
`"4!
`
`FIG. 2. Structures ol the pSV-neo hybrid plasmids. The pSV plasmids are composed of DNA segments from various sources:
`pBR322 DNA, represented by the solid black arc. contains the pBFl322 origin of DNA replication (pBFi322 ori) and the.3"3°‘3'"3"
`99'“ lN"P"lI "19 hatched segment represents the neo gene (1.4 kb fragment. Fig. 1); SV-40 DNA sequences are indicatedasm .
`dotted segments. The SV40 origin at DNA replication {SV40 ori} and the SV40 early promoter are present on a small tragmenl
`(SV40 map units 0.71—-0.65) immediately 5' to the neo segment. The plasmids pSV3-neo and pSV5-neo were constructed from
`pSV2-neo by insertion of either an intact SV40 early region or an intact polyoma early region that contains a duplication ol the
`viral origin of replication {8}.
`
`encc may be related to the ability of pSV3
`plasmids to replicate in monkey cells and pSV5
`plasmids to replicate in mouse cells (25). A
`comparable result was previously noted in
`transformations with pSV3-gpt and pSV5-gpt
`recombinants (26).
`
`Stable Integration of neo DNA SeqlIB"°°5
`after Transfection into Cultured Cells
`
`The content and organization of nco DNA
`r
`a number 0
`sequences have been examined in
`The Ltlf 0'
`stable G418 transformed cell lines.
`
`l
`
`l
`
`TABLE 2. Trrrrrsformrrtion frequencies for recombimm!
`neo plasmids"
`
`Host cells
`
`Ltk'
`
`3T6
`
`TC?
`
`pSV2-neo
`pSV3-neo
`pSV5-neo
`pBR-neo
`
`---3 X 10"
`-3 X 10"‘
`--3 x 10“
`-5 X 10'“
`
`7 X 104
`1.5 x 10"
`4 X 10'‘
`--1 x 10"’
`
`1 X 10“
`6 X 10"
`2 X 10"
`No colonies isolated
`
`" Transfonnation frequency is expressed as the fraction of cells plated
`that produce viable colonies in selective medium following transfection
`with saturating levels of DNA.
`
`J’. Mol. Appl. Gcrret.. Vol. I, No. 4, I982
`
`Merck Ex. 1005, pg 9
`
`Merck Ex. 1005, pg 93
`
`

`
`TRANSFORMATION T0 ANTIBIOTIC RESISTANCE
`
`3 ii
`
`3T6 transformants arising from transfections
`with pSV2-neo, pSV3-neo, or pSV5—neo do not
`contain plasmid—related sequences (limits of
`detection 0.5 plasmid molecules per cell) in the
`low molecular weight DNA of a Hirt superna-
`tant (41). However, cell DNA obtained from
`G418-resistant transformants contained nucle-
`
`otide sequences homologous to the transfecting
`plasmid DNA. High molecular weight DNA
`from a representative set of pSV2-neo Ltkr
`transformants was cleaved with EcoRI restric-
`
`tion endonuclease, electrophoresed in an
`agarose gel, transferred to DBM paper, and
`then hybridized with radioactively labeled
`pSV2-neo DNA (Fig. 3). Since pSV2-neo DNA
`contains a single EcoRI restriction site, each
`integrated plasmid copy should produce two
`hands after hybridization with the pSV2-neo
`probe if there is no rearrangement or scram-
`bling of the plasmid DNA sequences during in-
`tegration. Most of the transformants that have
`been examined appear to have a low plasmid
`copy number, generally one to five copies per
`cell. Transformant 2G is exceptional in that
`there are multiple copies of the plasmid and a
`substantial amount of apparent linear pSV2-
`neo DNA is produced by digestion with EcoRI
`cndonuclease (Fig. 3). Since no circular plas-
`mid DNA was present in the Hirt supernatant
`fraction prepared from 2G cells and the high
`molecular weight DNA of the Hirt pellet hy-
`bridized strongly to a radioactively labeled
`pSV2-neo probe (results not shown), it is likely
`that transformant 2G contains either a tandem
`
`array of integrated plasmid DNA or autono-
`mously replicating polymeric plasmid DNA.
`Although the latter possibility can not be
`excluded,
`it seems less likely because the
`G418-resistant phenotype as well as the copy
`number and organization of the neo DNA se-
`quences do not change after passaging in non-
`selective medium (see below).
`The Ltk‘ 2F and 2G cell lines have been
`cultured for over 3 months (approximately 100
`cell generations) in medium lacking or con-
`taining G418 (200 pig/ml); such cells remain
`fully resistant to G418 when challenged in
`
`t.tk”l'pSV2neo
`f
`
`2F
`
`2G|2Jl2I(‘M
`
`20 Rb-'-v
`
`, “ .-v‘I
`
`I
`
`k-=— 3 ~
`
`-- -:.l.'3§:‘...‘:
`
`Q
`
`1 kl:-'-
`
`FIG. 3. Detection of pSV2-neo DNA sequences in trans-
`formed cells. High molecular weight cell DNA was extracted
`from representative G418-resistant Ltk‘ cell lines that had
`been transformed with pSV2-neo plasmid DNA. The DNA
`samples were digested with an excess of Ecofil restriction
`endonuclease and then fractionated by electrophoresis in a
`0.8% agarose gel. DNA in the gel was transferred to DBM
`paper and hybridized with radioactively labeled pSV2-neo
`DNA. There is a single recognition site in pSV2-neo DNA for
`EcoFll cleavage and the marker track {M} shows the position
`of pSV2-neo linear DNA. Approximate molecular sizes were
`derived from A DNA restriction fragments in an adjacent slot
`on the gel.
`
`selective medium. The relative plating effi-
`ciency of the 2F or 2G cell lines after seeding in
`the presence or absence of G418 is within ex-
`perimental variation, the same (ratio 0.6 to
`0.9:I.0). Moreover,
`there was no discernible
`variation in the organization of the pSV2-neo
`DNA in the high molecular weight cell DNA
`obtained from 2F and 2G cell DNAs recovered
`
`at different times during their growth in the
`presence or absence of G413 (Fig. 4). Thus, it
`appears that the amount and arrangement of
`the pSV2-neo plasmid DNA sequences are sta-
`
`J. Moi. Appl. Gerirr.. Vol. i. No. 4, I982
`
`Merck Ex. 1005, pg 94
`
`Merck Ex. 1005, pg 94
`
`

`
`Ltk‘ pSV2neo2G
`Ltk‘ pSV2neo2F
`To-\ 1:-—"‘——':\
`'\%
`Ne’
`/
`'3'
`5."?
`<
`cs“
`9“
`s-
`<3‘
`o
`i"'I’I"|"|"I’i
`
`Ltk‘pSV2neo2G
`I?’-*—j\
`5.45
`55%
`0
`<5
`0
`llxl/ll
`
`to
`
`lG
`
`l'
`l
`I‘
`
`'
`i
`
`I
`:
`
`354
`
`P. J’. so UTHERN AND P. BERG
`
`FIG. 4. Detection of psv2-neo DNA se-
`quences after extended passaging of
`transformed cells. High molecular weight
`chromosomal DNA was extracted from
`parallel cultures of the transformed cell
`lines 2F and 2G that had been maintained
`in the presence (+G41B) or absence
`{—G418] of (3418 {see text}. DNA samples
`were digested with EcoFll endonuclease
`and analyzed as described in Fig. 3. The
`tracks at the extremities of the gel
`[labeled 2F and 2G) contain samples oi
`the original DNA preparations (see Fig. 3}.
`The central track contains parental Ltk'
`cell DNA. The panel at the right at the tig-
`ure is a shorter exposure during au-
`toradlography oi the Ltk“ pSV2-neo 2G
`cell DNA.
`
`'
`
`_
`
`_
`'
`'-
`‘-*5‘-fill-9
`"H"“"""
`
`“ --.
`“""‘ **""‘t
`-
`
`o
`
`Wh-
`
`.
`
`
`
`seems that the neo marker is associatedwith i
`the cellular genome relatively early and IS re-
`tained in the absence of G418 selection.
`
`, l
`
`Expression of the neo Gene in pSV2-neo
`Transformed Cells
`
`‘
`
`'
`The expression of neo in G418-resistant
`transformants has been confirmed by the do '
`tection of neo mRNAs and phosphotransferase _
`protein. Cytoplasmic, poly(A)+ RNA was iscr
`1
`lated from pSV2-neo transformed Ltk“ cells
`‘
`(isolates F and G) and hybridized to an end-
`_
`labeled DNA probe specific for pSV2-neo 56‘
`:
`quences. In each sample, digestion with 51
`nuclease and electrophoresis of the protected
`DNA in an agarose gel 69,40) yielded a sing“ .
`fragment of 1500 bases (Fig. 5). The production
`of a 1500 base fragment is consistent with H15
`occurrence of an RNA that extends throughout
`the length of the neo segment and suggests that
`
`Merck Ex. 1005, pg 9
`
`
`
`ble during growth of these two transformed cell
`lines under selective or nonselective condi-
`tions.
`
`We sought to determine if stable transforma-
`tion for the transduced marker could occur in
`the absence of selection. Accordingly, semi-
`conflucnt cultures of mouse 3T6 cells were
`transfected with pSV3-neo or pSV5—neo DNA
`and one set of cells was maintained under the
`standard conditions for transformation and an-
`
`other set was propagated in the absence of
`G418 for about 12 generations (2 weeks) and
`then the G418 selection was applied. The
`transformation frequency in the first set was
`10“ to 10'5 whereas the second set yielded
`about one-fifth the number of transformants
`
`predicted from the initial value and the number
`of cell divisions that had intervened before the
`
`selection was applied. Assuming that trans-
`formed and normal cells divide at the same rate
`
`in medium lacking G418 and considering the
`inherent inaccuracies of the experiment,
`it
`
`J. Mol. Appl‘. Gwen. Vol. I. No. 4, I982
`
`Merck Ex. 1005, pg 95
`
`

`
`TRANSFORMATION T0 ANTIBIOTJC RESISTA NCE
`
`335
`
`'
`s
`
`a
`
`3
`
`MARKERS
`/—|-I-/\-I-I-\
`
`=
`e
`
`"-.2
`3
`s e
`
`‘
`
`— PROBE
`
`....
`
`Q__-_
`
`b
`
`P
`
`T
`
`PROBE FOR pSV2-neo RNA
`
`EcoFIl
`
`Pvt.-‘ll 3”” Hmdulflgfll
`
`Baml-ll
`
`FIG. 5. Detection of neo sequences in cyto-
`plasmic FINA extracted from pSV2-neo trans-
`tcrmed Ltk' cetls. A DNA probe was prepared by
`5' end-labeling with “P at the BamHI restriction
`site at the 3' extremity ot the neo DNA lragment.
`(The BamHI site was not preserved in the con-
`struction of pSV2-neo and the probe for this ex-
`periment was prepared from another derivative
`plasmid.) Relevant restriction endonuclease
`cleavage sites are indicated in the diagram.
`FINA-DNA hybrids were formed under condi-
`tions ot DNA excess {40} and the S1 nuclease
`digests were fractionated on a 1% neutral
`agarose gel. Marker bands were produced by
`digesting the DNA probe with Bgll, Hindlll and
`Bglll restriction endonucleases. The DNA trag-
`rnent protected by the transformed cell RNAs is
`shown in the diagram as a continuous line
`under the representation of the DNA probe.
`
`’
`
`1110
`320
`290 72
`
`
`oessnveo PROTECTED
`FRAGM ENT
`
`1500
`
`the 5'-end of the RNA is located beyond the
`HindIII site, most probably in proximity to the
`SV40 early promoter. Note that although the
`CODY number of pSV2-neo DNA sequences is
`much higher in the 2G cells (Figs. 3 and 4), the
`amount of neo mRNA in the 2F and 2G trans-
`formed cell lines is about the same.
`To determine if the G413-resistant transfor-
`
`mants produce phosphotransferase protein,
`Semiconfluent cultures of normal and trans-
`l0l'med cells were labeled for 14 h at 37°C with
`l“H]leucine. Labeled extracts of E. coir‘ har-
`b°"i“3 pBR—neo or pBR322 were made for
`°°mDarison (see Materials and Methods). Im-
`mime complexes prepared from each of these
`extracts were electrophoresed in sodium dode-
`Clri sulphate polyacrylamide gels (Fig. 6). It is
`evident that the immunoprecipitate from E. coir’
`°?"1'YiI18 DBR-neo contains a peptide of 25
`kilodaltons, whereas the extract from cells
`lacking neo does not contain that peptide. The
`
`molecular weight estimate for AP!-i(3')Il ob-
`tained from E. coir’ agrees with earlier findings
`(48). Comparisons of the electrophoretic pat-
`tern of immune complexes obtained from un-
`transformed and two pSV2-neo transformed
`Ltk‘ cells (isolates F and G) reveal that the
`transformants contain a specific band corre-
`sponding to a protein of 28 kilodaltons with no
`indications of a protein of 25 kilodaltons. This
`difference in apparent molecular weight be-
`tween the immunoprecipitated APH(3’)1I
`found in E. cab‘ and the transformed mouse
`
`cells is considered in the Discussion section.
`
`Although the copy number of neo DNA is con-
`siderably greater in isolate 2G than 2F,
`the
`amount of phosphotransferase peptide appears
`to be about the same, even somewhat greater in
`2F; the same disparity occurs between neo
`DNA copy number and neo mRNA (see Fig.
`5). This suggests that many of the neo genes in
`isolate 2G may not be expressed.
`
`J. Moi. Appl. Gem:t.. Vol. I. No. 4. I982
`
`Merck Ex. 1005, pg 96
`
`Merck Ex. 1005, pg 96
`
`

`
`-u:_...__-:a-—-
`
`_|
`.
`
`336
`
`P. J’. SOUTHERNAND P. BERG
`
`lfl||‘tD‘-
`_
`
`-_.-
`
`.’ . ..
`
`i i
`
`i
`VP3 {21 Kdl ~;-»
`
`FIG. 6. Detection of phosphotranslerase
`by immunoprecipitation ol labeled cell
`extracts and S03 gel electrophoresis.
`The left panel shows the products from
`irnmunoprecipitation reactions with an-
`tiphosphotransferase antibody and ex-
`tracts ol E. coli cells containing the
`plasmids pBFl-neo or pBFl322. A unique
`band (indicated by the arrow} corre-
`sponding to a polypeptide of 25 l<iiodaI-
`tons (kd). is visible in the pBFl-neo ex-
`tract. The right panel shows the products
`from similar phosphotransferase im-
`munoprecipitation reactions using ex-
`tracts prepared lrom the parental Ltk‘ cell
`line and the G418-resistant Ltk' cell lines
`(2F and 2G). A specific band {indicated by
`the arrow] corresponding to a polypep-
`tide ol about 28 lrilodaltons is detected in
`the 2F and 2G transformed cell extracts.
`Possible explanations lor the apparent
`dilterence in molecular weight between
`the bacterial and the mammalian phos-
`phctransferase are in the Discussion. The
`molecular weights of the phosphotrans-
`lerase proteins have been estimated by
`rolerence to ‘H-labeled SV40 virion pro-
`toins.
`
`i4cxdl-{'-"'-_'-'-
`vr
`VP21l38.5Kdl ‘E? '-
`__
`'
` _b
`
` .
`
`" VP] [40 KG}
`-- VP2l33.5Klii
`
`__.n_na-Iii-nIi_
`
`were grown either in mycophenolic acid (MPM
`to select for gpt transformants,
`in G418 to
`select for neo transformants, or in medium
`containing both MPA and G418 to select di-
`l
`rectly for the double transformant.
`lndepen-
`dent colonies of each type were collected and *
`expanded to approximately 5 X 10“ cells with -
`the original selection conditions. Cells were I
`then" tested for the expression of the non-
`selected marker gene by plating in the opposite '
`
`pBR322 ori
`
`Amp“
`
`
`
`,
`.
`I
`
`N l
`
`FIG. 7. Structure of pSV2-neo.-Svgpt recombinant plasmid.
`This plasmid contains two SWO hybrid early region Iran»
`scription units to direct the expression of the gpt and neo
`marker genes independently (25). The component sequences
`can be identified according to the legend in Fig. 2.
`
`Merck Ex. 1005, pg 9
`
`Maintenance and Expression of a
`Nonselected Gene in Transformed Cells
`
`One objective in seeking vectors with domi-
`nant selectable markers is to facilitate the in-
`troduction and maintenance of genes that do
`not confer a selection. To determine if neo
`
`could be used for this purpose, cells were
`transfected with DNAs containing the neo and
`gpt markers, selection was applied for one or
`the other (or both) of the genes and transfor-
`mants were scored for expression of the non-
`selected marker. In this experiment the neo
`and gpt transcription units from the pSV2-
`plasmids were linked in a double marker plas-
`mid (pSV2-neo-Si/‘gpt, Fig. 7) (25). For com-
`parison, and to obtain information on cotrans-
`duction frequencies with