on May 10, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`G. Urlaub, L. Chasin, Pror. Natl. Acad. Sci.
`U.S.A. 76, 1373 (1979).
`9. L. H. Gra, G. Urlaub, L. Chasin, Somat. Cell
`Genet. 5, 1031(1979).
`10. S. C. Lester, S. K. LeVan, C. Steglich, R. De-
`Mars, ibid. 6, 241 (1980).
`11. M. Perucho, D. Hanahan, L. Lipsich, M. Wig-
`ler, Nature (London) 285, 207 (1980).
`12. 1. Lowy, A. Pellicer, J. Jackson, G. K. Sim, S.
`Silverstein, R. Axel, in preparation.
`13. J. Littlefield. Proc. Natl. Acad. Sci. U.S.A. 50,
`568 (1963).
`14. W. Szybalski, E. H. Szybalska, G. Ragni, Nail.
`Cancer Inst. Monogr. 7, 75 (1962).
`15. S. Kit, D. Dubbs-, L. Piekarski, T. Hsu, Exp.
`Cell Res. 31, 297 (1963).
`16. W. H. Lewis, P. R. Srinivasan, N. Stokoe,
`L. Siminovitch, Somat. Cell Genet. 6, 333
`(1980).
`17. M. Wigler, M. Perucho, D. Kurtz, S. Dana, A.
`Pellicer, R. Axel, S. Silverstein, Proc. Natl.
`Acad. Sci. U.S.A. 77, 3567 (1980).
`18. F. L. Graham and A. J. van der Eb, Virology 52,
`456 (1973).
`19. M. Wigler, R. Sweet, G. K. Sim, B. Wold, A.
`Pellicer, E. Lacy, T. Maniatis, S. Silverstein, R.
`Axel, Cell 16, 777 (1979).
`20. J. Spizizen, B. E. Reilly, A. H. Evans, Annu.
`Rev. Microbiol. 20, 321 (1966).
`21. P. J. Kretschmer, A. C. Y. Chang, S. N. Cohen,
`J. Bacteriol. 124, 225 (1975).
`22. E. M. Southern, J. Mol. Biol. 98, 503 (1975).
`23. J. C. Fiddes, P. H. Seeburg, F. M. DeNoto,
`R. A. Hallewell, J. D. Baxter, H. M. Good-
`man, Proc. Natl. Acad. Sci. U.S.A. 76, 4294
`(1979).
`24. K. Yamamoto, V. L. Chandler, J. Riug, D. Uck-
`er, Cold Spring Harbor Symp. Quant. Biol., in
`press.
`
`25. W. F. Flintoff, S. V. Davidson, L. Siminovitch,
`Somat. Cell Genet. 2, 245 (1976).
`26. See R. Mulligan and P. Berg, Science 2W9, 1422
`(1980).
`27. A. Peilicer, M. Wigler, R. Axel, S. Silverstein,
`Cell 14, 133 (1978).
`28. M. Perucho, D. Hanahan, M. Wigler, Cell, in
`press.
`29. D. Robins, A. Henderson, S. Ripley, R. Axel, in
`preparation.
`30. J. Smiley, D. A. Steege, D. K. Juricek, W. D.
`Summers, F. H. Ruddle, Cell 15, 455 (1978).
`31. M. Perucho and M. Wigler, in preparation.
`32. D. Hanahan, D. Lane, L. Lipsich, M. Wigler,
`M. Botchan, Cell 21, 127 (1980).
`33. B. Wold, M. Wigler, E. Lacy, T. Maniatis, S.
`Silverstein, R. Axel, Proc. Natl. Acad. Sci.
`U.S.A. 76, 5694 (1979).
`34. N. Mantei, W. Boll, C. Weissman, Nature (Lon-
`don) 281, 40 (1979).
`35. R. C. Mulligan, B. H. Howard, P. Berg, ibid.
`277, 108 (1979).
`36. D. Hamer and P. Leder, ibid. 281, 35 (1979).
`37. C. Weissman, personal communication.
`38. E. C. Lai, S. L. Woo, M. E. Bordelon-Riser, T.
`H. Fraser, B. W. O'Malley, Proc. Natl. Acad.
`Sci. U.S.A. 77, 244 (1980).
`39. R. Breathnach, N. Mantei, P. Chambon, ibid.,
`p. 740.
`40. P. N. Goodfellow, E. A. Jones, V. van Hegnin-
`gen, E. Solomon, M. Bobrow, V. Miggiano, W.
`F. Bodmer, Nature (London) 254, 267 (1975).
`41. Z. Schlegel and T. L. Benjamin, Cell 14, 587
`(1978).
`42. T. Maniatis, R. Hardison, E. Lacy, J. Lauer, C.
`O'Connell, D. Quon, G. K. Sim, A. Efstratiadis,
`ibid. 15, 687 (1978).
`43. W. D. Benton and R. W. Davis, Science 196,
`180 (1977).
`
`44. M. Grunstein and D. S. Hogness, Proc. Natl.
`Acad. Sci. U.S.A. 72, 3961 (1975).
`45. G. A. Galan, W. H. Klein, R. J. Britten, E. H.
`Davidson, Arch. Biochen". Biophys. 179, 584
`(1977).
`46. M. Wigler, R. Sweet, G. K. Sim, B. Wold, A.
`Pellicer, E. Lacy, T. Maniatis. S. Silverstein. R.
`Axel, in Eucaryotic Gene Regulation, R. Axel,
`T. Maniatis, C. F. Fox, Eds. (Academic Press,
`New York, 1979), p. 457.
`47. For review see T. Maniatis, E. F. Fritsch, J.
`Lauer, R. M. Lawn,. Annu. Rev. Genet., in
`press.
`48. B. Steinberg, R. Pollack, W. Topp, M. Botchan,
`Cell 13, 19 (1978).
`49. For a discussion of this problem for endogenous
`cellular genes see L. Siminovitch, in Eucaryotic
`Gene Regulation, R. Axel, T. Maniatis, C. F.
`Fox, Eds. (Academic Press, New York, 1979),
`pp. 433-443.
`50. R. Sweet, J. Jackson, A. Pellicer, M. Ostrander,
`S. Silverstein, R. Axel, in preparation.
`51. C. Waalmijk and R. A. Flavell, Nucleic Acids
`Res. 5, 4031 (1978).
`52. A. P. Bird, M. H. Taggard, B. A. Smith, Cell 17,
`889 (1979).
`53. J. L. Mandel and P. Chambon, Nucleic Acids
`Res. 7, 2081 (1980).
`54. A. H. Hinnen, J. B. Hicks, G. R. Fink, Proc.
`Natl. Acad. Sci. U.S.A. 75, 1929 (1978).
`55. R. DeMars, Mutat. Res. 24, 335 (1974).
`56. A. S. Henderson, M. T. Yu, K. C. Atwood,
`Cytogenet. Cell Genet. 21, 231 (1978).
`57. This work was supported by grants from the
`U.S. Public Health Service, National Institutes
`of Health, to R.A. (CA 16346 and CA 23767) and
`to S.S. (CA 17477).
`7 July 1980
`
`Expression of a Bacterial
`Gene in Mammalian Cells
`
`R. C. Mulligan* and P. Berg
`
`The most powerful tools now available
`for studying the molecular anatomy of
`eukaryote genes and chromosomes, par-
`ticularly those of higher vertebrates, are
`restriction enzymes,
`simple
`physical
`methods for separating and visualizing
`DNA molecules, molecular cloning, and
`rapid DNA sequencing. In just a few
`years, the application of these tech-
`niques has produced a qualitative change
`in our views of gene structure and organ-
`ization in mammalian and other et.
`karyote organisms. Such newly coined
`terms as gene libraries,
`split genes,
`pseudogenes, and transposons and pic-
`turesque references to "'shotgunning,"
`or even ""walking and jumping along
`
`chromosomes," reflect this revolution
`and enrich the lexicon of modern ge-
`netics. Solving the nucleotide sequence
`of the chromosomal locus encoding the
`entire human 8-globin-like gene cluster,
`a region encompassing about 65 kilobase
`pairs (kbp) (1), would have been consid-
`ered visionary 10 years ago, but now that
`prospect looms on the horizon as a fea-
`sible undertaking.
`Nevertheless, in spite of this impres-
`is worth considering
`sive progress it
`whether knowing the molecular anatomi-
`cal details of genes can, by itself, explain
`the subtleties of gene expression and reg-
`ulation during growth and development.
`Put another way, can we deduce the
`mechanisms of transcription initiation,
`splicing, and polyadenylation from the
`R. C. Mulligan is a predoctoral fellow and P. Berg
`is Willson Professor of Biochemistry in the Depart-
`nucleotide sequences of isolated genes?
`ment of Biochemistry, Stanford University Medical
`Can we expect to learn from the nucle-
`Center, Stanford, California 94305.
`0036-8075180/0919-1422$01.50/0
`Copyright X) 1980 AAAS
`
`1422
`
`otide sequence of the human y- and /-
`globin genes why the former is expressed
`only during fetal life and the latter only in
`adulthood? Most likely not; at least not -
`without an assay for the biological activi-
`ty of the genes in question.
`About 10 years ago we began to con-
`sider how the biological activity of isolat-
`ed genes could be assayed. That interest
`coincided with another preoccupation
`concerned with devising a virus-mediat-
`ed transducing system for cultured mam-
`malian cells. The overlap of these two in-
`terests culminated in a general approach
`for introducing isolated genes and their
`modified derivatives into the genomes of
`cultured mammalian cells (2). Our imme-
`diate goal was, and still is, to character-
`ize the physical state, expression, and
`regulation of the new genes in their
`transduced hosts.
`Because bacteriophages had proved to
`be so versatile for transducing genes in-
`to bacterial cells (3), simian virus 40
`(SV40), a mammalian virus, was adopted
`as the vector to mediate the gene trans-
`fer. SV40 was chosen because its mini-
`chromosome propagates vegetatively or
`becomes stably integrated into selected
`host cell genomes. SV40 was also at-
`tractive because its genes and their cor-
`responding functions had been identified
`and experiments were under way to map
`the genes to specific regions of the vi-
`rus's DNA. Subsequently, the entire
`5243-base pair (bp) sequence of the cir-
`cular viral DNA was solved (4, 5), and
`SCIENCE, VOL. 209, 19 SEPTEMBER 1980
`
`Merck Ex. 1002, pg 69
`
`

`
` on May 10, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`extensive information about the replica-
`tion, expression, and regulation of the
`viral genome in different cells has be-
`come available (6).
`To avoid a dependence on biologically
`generated transducing viruses and to in-
`crease the probability of obtaining spe-
`cific trafisducing genomes, we elected to
`construct, in vitro, recombinants be-
`tween SV40 DNA and the gene of inter-
`est. This was accomplished by ligating
`appropriate DNA fragments to whole (2)
`or subgenomic segments (7, 8) of viral
`DNA via enzymatically synthesized ho-
`mopolymeric cohesive termini (2,
`9).
`Today, ligation of two DNA molecules is
`usually accomplished via ends that are
`generated by restriction enzyme cleav-
`age of natural or engineered restriction
`sites (10).
`Initially,
`the recombinant genomes
`were propagated as virions and trans-
`duction occurred concomitantly with vi-
`rus infection (7, 8, 10). Since this experi-
`mental design requires that the recombi-
`nant genomes replicate, they must con-
`tain the origin of SV40 DNA replication
`(ori). Furthermore, to encapsidate the
`recombinant genome into a virion, the
`DNA molecule must be smaller than 5.3
`kbp, that is, about one mature viral DNA
`length. Since the vector lacks genetic
`functions coded by the excised DNA
`segment, the recombinant genomes are
`defective and, therefore, must be propa-
`gated with a helper virus that supplies
`the missing gene product or products. In
`our protocol (7, 8, 10) the recombinant
`genome retains at least one functioning
`virus gene, and consequently, it can
`complement a defective gene in the help-
`er virus. For example, recombinants in
`which the DNA insert replaces all or part
`of SV40's late region can be propagated
`with SV40 mutants that have a defective
`early region (for example, tsA mutants at
`high temperature); similarly, recombi-
`nants in which early region segments are
`replaced by the DNA implant can be
`propagated with a helper that is defective
`in its late region (for example, tsB mu-
`tants at high temperature).
`Our early attempts to obtain expres-
`sion of cloned segments as distinct mes-
`senger RNA's (mRNA) and proteins
`were negative (7, 8). Subsequently, re-
`combinant genomes containing either a
`1B-globin complementary DNA
`rabbit
`(cDNA) (11), a Drosophila melanogaster
`gene for histone H2b (12) or a cDNA
`coding for mouse dihydrofolate reduc-
`tase (DHFR) (13), in place of portions of
`SV40's late region, were constructed.
`After infection of monkey cells, each of
`the im-
`the recombinants expresses
`planted gene sequence as novel hybrid
`19 SEPTEMBER 1980
`
`mRNA's; moreover, the proteins 13-glo-
`bin (10), histone H2b (14), and mouse
`DHFR (15) are synthesized at levels
`comparable to those of SV40 late pro-
`teins. Similar successes in obtaining ex-
`pression of cloned genes have been
`achieved by Hamer and Leder with re-
`combinants carrying the mouse genomic
`,3-globin (16) or a-globin (17) genes, a re-
`sult that established proper splicing of
`the globin intervening sequences and
`translation of the resulting mRNA's in a
`heterologous host.
`
`E. coli XGPRT differs from the anal-
`ogous mammalian enzyme, hypoxan-
`thine-guanine phosphoribosyltransferase
`(HGPRT), in that xanthine is consid-
`erably more active as a substrate than
`hypoxanthine in nucleotide
`synthesis
`(19). By contrast, mammalian HGPRT
`does not utilize xanthine efficiently as a
`substrate; indeed, mammalian cells do
`not convert xanthine toxainthylic acid or
`to guanylic acid at a significant rate.
`In this article we describe the isolation
`ofEcogpt and its recombination with ap-
`
`Summary. Transfection of cultured monkey kidney cells with recombinant DNA
`constructed with a cloned Escherichia coli gene that codes for xanthine-guanine
`phosphoribosyltransferase and several different SV40 DNA-based vectors, results in
`the synthesis of readily measurable quantities of the bacterial enzyme. Moreover, the
`physiological defect in purine nucleotide synthesis characteristic of human Lesch-
`Nyhan cells can be overcome by the introduction of the bacterial gene into these cells.
`
`Until recently, our principal focus has
`been to exploit the ability of the recombi-
`nant genomes to replicate in the virus's
`permissive host. During vegetative mul-
`tiplication, the transduced genomes are
`amplified about 104- to 105-fold, thereby
`ensuring high yields of the gene prod-
`ucts. Such studies have been informative
`about the necessity and mechanistic sub-
`tleties of RNA splicing, the rules govern-
`ing expression of coding sequences in-
`serted at different loci in SV40 DNA, and
`about many facets of SV40 gene expres-
`sion itself. But this experimental design
`also has a distinct shortcoming: The
`cell is killed during the course of the
`infection, thereby precluding the oppor-
`tunity to monitor the transduced gene's
`expression in continuously multiplying
`cells. Moreover, our current collection
`of recombinants can only be studied in
`permissive cells, that is, those that can
`amplify the recombinant genomes. This
`constraint excludes many specialized
`and differentiated animal cells as hosts
`for the transduced genes.
`To circumvent these disadvantages,
`we sought to develop transducing vec-
`tors that could be introduced and main-
`tained in a variety of cells. Our initial at-
`tempts in this direction indicated that our
`approach would be facilitated by the
`availability of a gene whose function
`could be selected for. Since our experi-
`ence indicated that the proper position-
`ing of protein coding sequences in SV40
`vectors would ensure their expression in
`transduced cells (18), we explored the
`possibility of using a bacterial gene for
`that purpose. The gene chosen was the
`Escherichia coli gene (Ecogpt) coding
`for the enzyme xanthine-guanine phos-
`phoribosyltransferase
`(XGPRT).
`The
`
`propriate SV40 DNA based transducing
`vectors. Transfection of a variety of cul-
`tured mammalian cells with such re-
`combinant DNA's results in the forma-
`tion of readily measurable quantities of
`coli XGPRT. Moreover, human
`E.
`Lesch-Nyhan cells that have been trans-
`fected with vectors containing Ecogpt
`DNA grow under conditions in which the
`parental cells do not survive. This result
`indicates that E. coli XGPRT can over-
`come the Lesch-Nyhan cell's physiologi-
`cal defect in purine nucleotide synthesis,
`and suggests that Ecogpt may be a gener-
`ally useful selectable marker for mam-
`malian cells.
`
`Isolation of Ecogpt for Introduction into
`SV40 DNA Vectors
`Ecogpt was obtained by a series of ma-
`nipulations and subcloning operations
`(Fig. 1). The existence and availability of
`the transducing phage Xgpt (see legend to
`Fig. 1) greatly facilitated the isolation in
`that it provided a more enriched source
`of the gpt gene than E. coli DNA itself.
`Equally crucial to the success of the gene
`isolation was the availability of suitable
`Gpt- mutants ofE. coli (20) whose trans-
`formation to Gpt+ could be readily mon-
`itored. With one notable exception, the
`subcloning procedure outlined in Fig. 1
`employs the same strategy used in ob-
`taining Ecotdk, the gene coding for thy-
`midine kinase (8). In this instance, how-
`ever, the 5' and 3' ends of the gpt-con-
`taining DNA fragment has been modified
`to contain either a Hind III or a Bam HI
`cohesive end. As a result, ligation of the
`gpt fragment to the corresponding cohe-
`sive ends of the pBR322 plasmid and
`1423
`
`Merck Ex. 1002, pg 70
`
`

`
` on May 10, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`hours after infection with SVGT5-gpt(l)
`contain at least 40 times more XPRT ac-
`tivity than uninfected CVl extracts (in-
`set in Fig. 3A). Insignificant XPRT activ-
`ity was detected in comparable extracts
`from cells infected with SV40 or SVGT5.-
`gpt(e). Also, Fig. 3A shows that in cells
`infected with SVGT5-gpt(l) nearly half of
`the enzymatic activity that converts 14C-
`labeled guanine to 14C-labeled guanosine
`monophosphate (GMP) resembles E. coli
`XGPRT in its insensitivity to inhibition
`by unlabeled hypoxanthine. By contrast,
`the same reaction with uninfected cell
`extracts or extracts obtained after infec-
`tions with either SV40 or SVGT5-gpt(e)
`is completely inhibited by unlabeled
`hypoxanthine, a property characteristic
`of the cellular HGPRT.
`Electrophoretic evidence also sup-
`ports the contention that E. coli XGPRT
`is formed after infection with SVGT5-
`gpt(l) (Fig. 3B). After polyacrylamide gel
`electrophoresis of extracts from cells in-
`fected with SV40, SVGT5-gpt(e), and
`SVGT5-gpt(l) or uninfected cells, assays
`of GPRT activity in situ show that only
`SVGT5-gpt(l) induces an enzyme activi-
`ty that comigrates with E. coli XGPRT.
`Each of the other extracts contain mon-
`key cell HGPRT but no detectable E.
`coli XGPRT. From these experiments,;
`we infer that the new enzymatic activity
`is E. coli XGPRT. Note that SVGT5-
`gpt(e) does not produce XGPRT. This
`might reflect either a failure to transcribe
`the gpt sequence from the E. coli pro-
`moter or an inability to process that tran-
`script. More likely, however, the gpt se-
`quence is expressed only from SV40's
`promoter, and the transcript is processed
`by means of SV40 splicing signals. Thus,
`mammalian cells can express the bacte-
`rial gene provided its coding sequence is
`properly positioned in a suitable vector.
`An analysis of the precise structure of
`the mRNA's made by SVGT5-gpt as
`well as other recombinants should en-
`able us to ascertain what features of the
`recombinant structures are significant
`for obtaining expression of the implanted
`DNA segments.
`
`New Vectors for Transducing
`Genes into Cells
`As indicated earlier, a primary objec-
`tive of this work is the development of
`vectors that can be used to introduce and
`maintain genes of interest in mammalian
`cells. Such vectors could be integrated
`into the cell's chromosome or be main-
`tained as part of an autonomously repli-
`cating DNA molecule. If such vectors
`SCIENCE, VOL. 2OW
`
`SV40 vectors is more efficient, and iso-
`meric recombinants, which differ only in
`the orientation of the Gpt relative to the
`vector, are equally frequent products.
`Since both orientational isomers of
`pBR322-gpt were equally efficient
`in
`complementing the Gpt- defect of the
`transformed E. coli host, and since the
`same amounts of XGPRT were induced
`in the transformants, we surmise that the
`l-kbp Ecogpt DNA contains all the ge-
`netic signals needed to express XGPRT,
`that is, the transcriptional promoter and
`terminator as well as the protein coding
`sequence.
`
`Expression ofEcogpt in
`Cultured Monkey Cells
`We have already described (10) an
`SV40 vector (SVGT5) which lacks the
`DNA segment between map position
`0.945 and 0.145. Vector SVGT5 still re-
`tains the putative SV40 late promoter
`and polyadenylation sites, as well as the
`late mRNA leader sequences and splice
`junctions; it lacks most of the region
`
`specifying the major virus capsid pro-
`tein. Since SVGT5 was active in promot-
`ing the expression of cDNA's, the
`cloned Ecogpt DNA was ligated to
`SVGT5 via the complementary Hind III
`and Bam HI cohesive ends and the re-
`sulting recombinant, SVGT5-gpt, was
`cloned in monkey cells as previously de-
`scribed (10). As expected, two SVGT5-
`gpt derivatives were obtained, differing
`only in the orientation of the gpt seg-
`ment. Figure 2 shows the structure of
`SVGT5-gpt, emphasizing the location of
`the gpt segment (the solid black seg-
`ment). The two orientational isomers of
`the SVGT5-gpt are referred to hereafter
`as SVGT5-gpt(l) and SVGT5-gpt(e) to
`indicate that one recombinant has the
`gpt coding sequence linked to SV40's
`late strand (1) and the other to the early
`strand (e), respectively. The structures
`of the putative gpt mRNA's are also
`shown to emphasize the significance of
`vector DNA sequences that specify the
`5' and 3' ends and the splice junctions of
`the predominant SV40 mRNA families
`(21, 22).
`Extracts of CV I cells obtained 36 to 48
`
`6kbp
`E.coli
`
`Shear
`
`gpt
`
`dA/dT, tails, clone in pMB9
`
`Fig. 1. Scheme for isolation of a small DNA
`fragment containing the E. coli gpt gene. The
`DNA used for the isolation of Ecogpt was a
`phage containing an insertion of 6 kbp of E.
`coli DNA spanning the pro gpt lac region of
`the E. coli chromosome (35). This phage, Xgpt,
`efficiently
`a variety
`transduces
`of Gpt-
`strains, including pro gpt lac deletion strains,
`to Gpt+. Xgpt DNA was sheared to an average
`size of 2 kbp and short dT (deoxythymidylate)
`chains were added to the termini of the
`sheared fragments as described (2, 7-9). The
`poly(dT) tailed phage DNA was annealed with
`poly(dA) tailed Eco RI endonuclease-cleaved
`pMB9 DNA. Then the recombinant molecules
`were transfected into E. coli JW2 (19), a
`Gpt- purE strain (the purE mutation prevents
`de novo synthesis of purine nucleotides). Tet-
`racycline-resistant (tel ) clones were screened
`for Gpt+ colonies by their ability to grow on a
`minimal salts medium (M9) containing glu-
`cose, vitamin-free casamino acids, thiamine,
`tryptophan, and 60 ,&g of guanine per millili-
`ter. Plasmid DNA isolated from one such col-
`ony was found by restriction endonuclease
`analysis to contain a 1.6-kbp segment of E.
`l
`coli DNA. Transfection of JW2 and E. coli
`* Ligate to SVGT vectors
`GP120-a Gpt- deletion strain (36)-with this plasmid DNA (pPT-l) yielded tet' transformants
`that were all Gpt+. Recombinant plasmids containing the Ecogpt segment in the opposite orien-
`tation were also obtained and these transformants were also Gpt+. The XGPRT activity in
`extracts from Gpt- or wild-type strains carrying pPT-l DNA was 20-fold higher than that found
`in wild-type E. coli strains. To obtain a smaller segment encoding the Ecogpt gene, pPT-l DNA
`was cleaved with Bgl I and Hinc II endonucleases, and the l-kbp fragment from within the
`Ecogpt insert was isolated by gel electrophoresis. The Bgl I endonuclease generated end of this
`frgment was made blunt by treatment with E. coli DNA polymerase and the four deoxyribonu-
`cleoside triphosphates (37). A mixture of decanucleotide linkers containing Hind III and
`Bam HI endonuclease recognition sequences (Collaborative Research) (each in 40-fold molar
`excess) were ligated to the blunt ends of the Ecogpt fragment with T4 DNA ligase (10). After
`restriction endonuclease digestion with both Bam HI and Hind III endonucleases, the fragment
`was ligated to Hind III and Bam HI cleaved pBR322 DNA with T4 DNA ligase (10). Trans-
`fection of the Gpt- strain of E. coli with the hybrid DNA molecules yielded Gpt+ clones that
`contained both orientational isomers of the Ecogpt insert.
`1424
`
`lTransfect E. coi gp
`Select E.cofi gpt
`Select Ecoli gpt+
`Bglll
`J
`
`Hinci
`
`egiI
`\
`
`(ct
`
`Cleave with Bgl and Hinc II
`Attach Hindll, BamHI linkers
`
`clone into PBR322 at
`Hindlll-BamHl sites
`Transfect E.coli gpt-
`Bgl II
`HindlIl
`Select E.coli gpt
`'3a Ham Hi Hind Bgj 1BamHl
`( 7)
`
`Cleave with Hind Ill, BamHl
`
`Merck Ex. 1002, pg 71
`
`

`
` on May 10, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`contained a selectable genetic marker,
`retention of the vectors and any non-
`selectable genes associated with them
`would be more likely and readily mon-
`itored. To accommodate genes of inter-
`est, the vectors should contain suitably
`positioned restriction sites for cloning
`appropriate DNA fragments. Last, but
`not least, considerable economies of
`time and money would be achieved if the
`manipulations needed to construct the
`desired recombinants could be per-
`formed in bacteria rather than cultured
`mammalian cells.
`The general features of the construc-
`tion and properties of the first generation
`of such vectors are shown in Fig. 4 (23).
`Each vector DNA can propagate as a
`plasmid in E. coli, hence the prefix p be-
`fore their designations. Both pSVlGT5
`and pSVlGT7 contain the SV40 DNA
`vectors SVGT5 and SVGT7 (18) shown
`here with the Ecogpt segment inserted at
`the nominal locations between SV40
`map positions 0.945 or 0.86 and 0.14,
`respectively. Joined to the SVGT5 or
`SVGT7 DNA is pBR322 DNA contain-
`ing a segment of DNA from SV40's early
`region (map position 0.14 to 0.325), liga-
`tion having occurred at their common
`Pst I restriction site at SV40 map posi-
`tion 0.28. This construction is conve-
`nient because, after preparation in E.
`coli, the SVGT5 or SVGT7 recombinant
`can be excised with Pst I endonuclease
`and propagated as a virus chromosome
`by transfection into monkey cells with
`the appropriate helper. Since pSVlGT5
`or pSV 1 GT7 contains an intact early re-
`gion, these recombinants can also repli-
`cate in primate cells, the permissive host
`for SV40. Moreover, because of the ap-
`proximately l-kbp duplication of SV40
`DNA sequences in the pSVlGT mole-
`cules (the segment between map posi-
`tions 0.145 and 0.325), homologous re-
`combination during replication generates
`the respective SVGT5 or SVGT7 re-
`combinants in situ.
`- Shown here with an Ecogpt segment,
`pSV2 consists of a 2.3-kbp portion of
`pBR322 DNA containing the plasmid's
`origin of DNA replication and ampicilli-
`nase gene joined to a segment of SV40
`DNA containing the SV40 early pro-
`moter upstream of Ecogpt; the Ecogpt
`segment is followed by another SV40
`DNA segment containing the small t an-
`tigen intervening sequence (24), the early
`region polyadenylation site and the 3'
`terminal segment of SV40's late region.
`Since pSV2-gpt DNA lacks the gene
`coding for SV40 T antigen, it does not
`replicate in monkey cells (25) although it
`can be propagated in E. coli.
`19 SEPTEMBER 1980
`
`(pSV3) or (ii) polyoma's early region
`with two origins of DNA replication (26)
`(pSV5). Since pSV3 and pSV5 vectors
`contain viral genes that promote DNA
`replication at their respective origins,
`pSV3 recombinants replicate in monkey
`cells and pSV5 derivatives replicate in
`mouse cells (25).
`
`Synthesis ofE. coli XGPRT in CV1 Cells
`Transfected with pSV-Ecogpt DNA's
`
`CVl cells were transfected with either
`pSVlGT5-gpt, pSVIGT7-gpt, pSV2-gpt,
`pSV3-gpt, or pSV5-gpt DNA (see legend
`to Fig. 5); after 72 hours cell extracts
`were subjected to electrophoresis in
`polyacrylamide slab gels and assayed in
`situ for GPRT (see legend to Fig. 3B).
`Each recombinant induced a GPRT ac-
`mobility
`tivity whose electrophoretic
`was characteristic of the E. coli enzyme
`(Fig. 5). Since only about 5 to 10 percent
`of the CVI cells are transfected by this
`procedure (27), the amount of E. coli
`GPRT activity is correspondingly lower
`relative to the amount of cellular GPRT
`and other proteins in the extract (see Fig.
`3B for comparison of the results of infec-
`tions with virus).
`It should be noted that GPRT activity
`can be detected in DNA transfections
`
`Fig. 2. Structure of SVGT5-gpt recombinant
`and expected structure of the hybrid SV40-
`gpt mRNA's. The pBR322-Ecogpt recombi-
`nant DNA's containing the 1-kbp Ecogpt seg-
`ment in both orientations (see Fig. 1) were
`cleaved with Hind III and Bam HI endo-
`nucleases to obtain the cloned inserts. These
`were ligated to SVGT5 DNA and propagated
`as viruses in CVI cells as described (10). The
`circle represents retained SV40 DNA, the
`dotted portion spans the deleted SV40 late
`segment, and the solid black segment within
`the deleted region is the l-kbp Ecogpt DNA
`fragment in either of the two orientations.
`Two size classes of hybrid SV40-gpt mRNA's
`are expected. The smaller and larger mRNA's
`have leaders characteristic of SV40 16S and
`19S mRNA's, respectively. The location of
`the 5' ends of the leaders are left unspecified
`because of the multiplicity of 5' ends (21, 38).
`The wavy lines represent regions that are
`spliced out of the mRNA's.
`
`Plasmids pSV3 and pSV5 also shown
`here with the Ecogpt DNA segment are
`derivatives of pSV2 that contain at
`pSV2's single Bam HI restriction site
`DNA segments which include either (i) a
`complete early region from SV40 DNA
`
`Extract
`
`XPRT activity
`(nmole/mir-mg)
`<0.l
`CV-1
`\CV-1 /SV40
`<30.1
`-y nCV-I/SVGT5gpt(e) <o.1
`3-4
`~~CV- 1 /SVGT5-gpt(1)
`
`0
`
`_
`
`0
`
`0
`
`_
`
`0
`
`a
`
`o
`c
`>
`ci,
`
`v
`0
`0
`2
`
`_6t\
`
`o 5e
`0\
`
`_ \ \
`
`E
`(L 2\
`
`e-0
`
`s
`
`W-GPR
`
`Monkey
`HGPRT
`
`Ecoli
`
`o GT5-gpt(e)____
`400
`160
`240
`320
`Hypoxanthine ("m)
`Fig. 3. E. coli XGPRT in monkey cells infect-
`SVGT5-gpt.
`ed
`with
`designations
`The
`SVGT5-gpt(e) and SVGT5-gpt(l)
`ex-
`are
`plained in the text. (Inset) XPRT was mea-
`sured by the conversion of "4C-labeled xan-
`thine (X) to "4C-labeled XMP as follows. Reaction mixtures (50 ,ul) contained 100 mM tris,
`pH 7.5, 4 mM MgCl2, bovine serum albumin (2 mg/ml), 2 mM phosphoribosylpyrophosphate
`and 180 pM 14C-labeled xanthine (58 mCi/mmole; Schwarz/Mann). After 15 minutes at 37°C, a
`portion (5 ,ul) of the reaction was mixed with unlabeled XMP and applied to cellulose thin-layer
`plates (Polygram CEL 300, Brinkmann Instruments) and chromatographed in IM ammonium
`acetate until the solvent had moved 8 cm. After the plates were dried, the area containing XMP
`was located under ultraviolet light, cut out, and counted. (A) The activity of the GPRT induced
`by SVGT5-gpt plotted as a function of the concentration of hypoxanthin. The GPRT activity in
`infected and uninfected CVI extracts was assayed by the conversion of 14C-labeled guanine to
`I4C-labeled GMP; the reactions were as described above, except that 20 iAM "4C-labeled guanine
`was used. After the incubation, the reaction mixture was spotted on DE81 filter disks, washed
`with 10 mM tris, pH 7.4, and counted. (B) Separation of monkey and E. coli purine phospho-
`ribosyltransferase activities by polyacrylamide gel electrophoresis and assay of the two en-
`zymes in situ. Extracts of SVGT5-gpt infected cells were subjected to electrophoresis on a 7.5
`percent polyacrylamide gel containing 0.2M tris, pH 8.5; GPRT activity was assayed in situ as
`described (38). The 3H-labeled GMP formed was detected by fluorography and exposure to x-
`ray film for 24 hours. The locations of E. coli and monkey GPRT are indicated by arrows.
`1425
`
`Merck Ex. 1002, pg 72
`
`

`
` on May 10, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`with pSV2-gpt DNA, a molecule that
`cannot replicate in monkey cells. Clear-
`ly, additional experiments are needed to
`relate the efficiency of gpt expression to
`the number ofgpt gene copies and to the
`extent of autologous repression that T
`antigen exerts on gpt transcription from
`the SV40 promoter (28).
`
`Transformation of Human Lesch-Nyhan
`Cells with pSV2-gpt and pSV3-gpt DNA's
`
`lack
`cells
`Lesch-Nyhan
`Human
`HGPRT (29); HGPRT-negative cells, un-
`like normal ones, cannot grow in a cul-
`ture medium containing hypoxanthine,
`aminopterin, and thymidine (HAT medi-
`um) (30). Therefore, it was of interest to
`know whether E. coli XGPRT would en-
`able Lesch-Nyhan cells to survive and
`multiply in HAT medium. In order to
`-test this point, skin fibroblasts, biopsied
`from a Lesch-Nyhan patient and subse-
`quently immortalized by SV40 transfor-
`mation
`(31), were transfected
`with
`pSV2-gpt or pSV3-gpt DNA's, as de-
`scribed in the legend to Fig. 3; trans-
`formant
`ere cultured and recovered
`as in Fig. 6.
`Cells that had not received DNA failed
`to survive in HAT medium, and no
`clones were detected after
`15 days.
`However, in cultures transfected with
`pSV2-gpt or pSV3-gpt DNA, there were
`detectable colonies within 7 to 10 days,
`and these were picked and subcultured
`after 15 days. Approximately 10 to 20
`colonies per 105 cells were obtained in
`the two sets of transfections, indicating a
`transformation frequency, under these
`conditions, of approximately 10-4. Two
`isolates each from the pSV2-gpt and
`pSV3-gpt transformed cells were tested
`for the presence of E. coli XGPRT after
`approximately 40 cell divisions in the se-
`lective medium (Fig. 6). Quite clearly,
`both sets of transformants contain the
`bacterial form of GPRT and, as expect-
`ed, lack the human GPRT.
`pSV2-gpt and pSV3-gpt transformed
`Lesch-Nyhan cells have also been sub-
`cultured in normal (nonselective) medi-
`um for about 12 generations. Extracts of
`such cells show no detectable diminution
`of the amount of bacterial GPRT. Al-
`though more extensive tests for the con-
`tinued maintenance and expression of
`Ecogpt are needed, the preliminary in-
`dications are that the transformants are
`genetically stable; that is, segregation of
`Ecogpt is infrequent. The copy number
`and physical state of the Ecogpt DNA in
`these transformants have not yet been
`characterized.
`
`SCIENCE, VOL. 209
`
`Eco-gpt
`
`0.86 (GT7)
`
`0.945 (GT5)
`
`0.14
`
`0.17
`
`0.28
`0.325
`
`AmpR
`
`RI
`
`Pstl
`
`pSV2-gpt
`11BamHl-
`
`PvuII3
`
`SV40 or
`
`HindIII
`
`Bg//l
`
`Eco-gpt
`
`AmpR
`
`pSVlGT5-gpt
`LATELAE pSVIGT7 9pt
`
`pBR322
`
`pBR322 ori
`
`ori
`
`.I EARLY"
`
`0.4
`
`0.17
`
`0.28
`
`0.325
`
`AmpR
`
`pBR322 ori
`
`X
`PVUII3@(
`PVUiII
`SV40 Ori
`
`RI
`
`titi
`
`/
`
`0.17
`
`0.
`
`pBR322 ori
`
`RI
`
`~0.17
`
`PstrI
`
`//
`
`pSV3-gpt
`
`BamHl
`
`<
`
`SV40-Tag
`
`IIX r
`
`BgI
`5V4Oori
`
`H,ndIII
`
`\\
`
`~~~~~
`
`~~ori
`
`PVUI3
`
`SV40 Or
`
`PSV5-gpt
`181BamHl
`,Ind
`
`//
`<
`
`\\
`
`PY-Tag
`
`Eco-gpt
`
`Ecogopt
`Fig. 4. Structure of plasmid vectors. All the vectors are shown with the gpt segment hatched.
`The solid black segments in each of the diagrams represent pBR322 DNA sequences: 4 kbp in
`the pSVl vectors and 2.3 kbp in pSV2, -3, and -5. In each instance, the pBR322 DNA segment
`contains the origin of pBR322 DNA replication and the ampicillinase gene. In the