on June 20, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`beled at the 5' end were purified by Sephadex G-
`50 chromatography in a 0.3M sodium acetate,
`0.OlM tris-HCI, pH 7.9, 0.001M EDTA system,
`and concentrated by precipitation with ethanol.
`6. A. Maxam and W. Gilbert, Proc. Natl. Acad.
`Sci. U.S.A. 74, 560 (1977).
`7. Low molecular weight DNA fragments were re-
`solved by electrophoresis at < 10 volt/cm in gels
`containing 10 percent polyacrylamide (acryla-
`N,N'-methylenebisacrylamide = 25: 1)
`mide:
`in one-half strength TBE [0.05M tris-borate, pH
`8.3, 0.001M EDTA, A. Peacock and C. Ding-
`man, Biochemistry 7, 668 (1968)] with 25 percent
`glycerol. Glycerol improves band sharpness dra-
`matically.
`8. J. Dahlberg and F. Blattner, in Virus Research,
`C. Fox and W. Robinson, Eds. (Academic
`Press, New York, 1973), p. 533.
`9. M. Rosenberg, personal communication.
`10. D. Kleid, Z. Humayun, A. Jeffrey, M. Ptashne,
`Proc. Natl. Acad. Sci. U.S.A. 73, 293 (1976).
`11. M. Rosenberg, B. de Crombrugghe, R. Musso,
`ibid., p. 717.
`12. G. Scherer, G. Hobom, H. Kossel, Nature
`(London) 256, 117 (1977).
`13. V. Pirotta, personal communication.
`14. E. Schwarz, G. Scherer, G. Hobom, H. K6ssel,
`personal communication.
`14a.G. Scherer, G. Hobom, H. Kossel, personal
`communication.
`15. B. Westmoreland, W. Szybalski, H. Ris, Sci-
`ence 163, 1343 (1969).
`16. D. Hewish, K. Smith, J. Spencer, Virology 74,
`363 (1976).
`17. A. Rambach, ibid. 54, 270 (1973).
`18. W. Dove, H. Inokuchi, W. Stevens, in The Bac-
`teriophage Lambda, A. D. Hershey, Ed. (Cold
`Spring Harbor Laboratory, Cold Spring Harbor,
`N.Y., 1971), p. 747.
`19. W. Szybalski et al., J. Cell. Physiol. 74 (Suppl.
`1), 33 (1969).
`20. H. Inokuchi, W. Dove, D. Freifelder, J. Mol.
`Biol. 74, 721 (1973).
`21. N. Kleckner, Virology 79, 174 (1977).
`22. M. Schnos and R. Inman, J. Mol. Biol. 55, 31
`(1971); A. Kozinski and B. Lindqvist, personal
`communication.
`23. The X ori region, defined by the ori- mutations,
`has two dPy - dPu stretches (18 bp and 18 of 21
`bp) in which the pyrimidines are on the r strand.
`The origin of Col El replication, defined by the
`site at which deoxyribonucleotides are added
`onto an RNA primer in vitro, is 20 bp away from
`a dPy - dPu stretch (21 of 24 bp) with purines on
`the transcribed strand [J.-I. Tomizawa, H. Oh-
`mori, R. Bird, Proc. Natl. Acad. Sci. U.S.A. 74,
`1865 (1976)]. The origin region for phage fd
`DNA replication, defined by an in virto RNA
`primer, contains a dPy * dPu stretch (29 of 32
`bp) with pyrimidines on the transcribed strand
`(C. Gray, R. Sommer, C. Polke, E. Beck, H.
`Schaller, personal communication). The pyrimi-
`dine tract CTC5 lies in the region of the A gene
`dependent gap in OX174 at the start of RF to RF
`replication [S. Eisenberg, B. Harbers, C. Hours,
`D. Denhardt, J. Mol. Biol. 99, 197 (1975)].
`24. F. Sanger, G. Air, B. Barrell, N. Brown, A.
`Coulson, J. Fiddes, C. Hutchison, P. Slocombe,
`M. Smith, Nature (London) 265, 687 (1977); P.
`Baas, H. Jansz, R. Sinsheimer, J. Mol. Biol.
`102, 633 (1976); S. Eisenberg, B. Harbers, C.
`liours, D. Denhardt, ibid. 99, 197 (1975).
`25. R. Fisher, The Genetical Theory of Natural Se-
`lection (Clarendon, Oxford, 1930).
`26. We believe that the deletions observed in the nu-
`cleotide sequences for r93, r99, and r96 corre-
`spond to the orti lesions. The ori- mutants were
`isolated without mutagenesis and occur with fre-
`quencies inconsistent with multiple mutations.
`The deletions could not have arisen during clon-
`ing, because they are present in the ori- phages
`ANam7NamS3cI8S7r93
`ANam7NarnS3c-
`and
`1857r99. The 0 protein made from the DNA of
`these phages in vitro is smaller than that made
`from DNA of the oril parent phage when mea-
`by SDS polyacrylamide
`sured
`gel
`elec-
`trophoresis (J. Yates and M. Furth, unpublished
`results, using techniques described in J. Yates,
`W. Gette, M. Furth, M. Nomura, Proc. Natl.
`Acad. Sci. U.S.A. 74, 689 (1977)]. The size of
`each deletion is consistent with the changes in
`molecular weights of the 0 protein made in vitro
`and of DNA restriction fragments. Finally, the
`recombinational map of the ori- mutations is
`consistent with the order of the sequenced dele-
`tions (r93-r99-r96), indicating that the deletions
`are responsible for the ori- pheonotype of these
`mutants (2).
`27. J. Szpirer, Mol. Gen. Genet. 114, 297 (1972); M.
`Furth, C. McLeester, W. Dove, in preparation;
`L. Lambert, R. Thomas, R. Monnat, W. Szy-
`balski, personal communication.
`28. By the criterion of heteroduplex pairing, the re-
`1056
`
`gion of DNA encoding the COOH-terminal half
`of the 0 protein appears to be lar%ely conserved
`in all five lamboid coliphages studied [M. Simon,
`R. Davis, N. Davidson, in The Bacteriophage
`Lambda, A. D. Hershey, Ed. (Cold Spring Har-
`bor Laboratory, Cold Spring Harbor, N.Y.,
`1971), p. 313; M. Fiandt, Z. Hradecna, H. Loze-
`ron, W. Szybalski,
`ibid.,
`p.
`329]. Com-
`plementation analysis shows that P gene func-
`tion can be provided to X by each of the other
`four lamboid phages [W. Dove, Annu. Rev.
`Genet. 2, 305 (1969); J. Szpirer and P. Brachet,
`Mol. Gen. Genet. 108, 78 (1970)]. In 4+80/A hy-
`brid phages such as Ximm80hy42 the 0 gene (27)
`and 0 protein iteself have been shown to be hy-
`brid (J. Yates and M. Furth, unpublished obser-
`vations). There is genetic evidence for direct in-
`teraction between 0 and P proteins [J.-I. Tomi-
`zawa, in The Bacteriophage Lambda, A. D.
`Hershey, Ed. (Cold Spring Harbor Laboratory,
`Cold Spring Harbor, N.Y., 1971), p. 549; M.
`Furth, C. McLeester, W. Dove, in preparation;
`(2)] and between P protein and the E. coli repli-
`cation apparatus [C. Georgopoulos and I. Her-
`skowitz, in The Bacteriophage Lambda, A. D.
`Hershey, Ed. (Cold Spring Harbor Laboratory,
`Cold Spring Harbor, N.Y., 1971), p. 553; H.
`Saito and H. Uchida, J. Mol. Biol.
`113, 1
`(1977)].
`29. S. Hayes and W. Szybalski, in DNA Synthesis
`
`and Its Regulation, M. Goulian, P. Hanawalt,
`C. Fox, Eds. (Benjamin, Menlo Park, Calif.,
`1975), p. 486.
`30. B. Williams, M. Furth, K. Kruger, D. Moore,
`K. Denniston-Thompson, F. Blattner, in prepa-
`ration.
`30a.W. Dove, E. Hargrove, M. Ohashi, F. Haugii,
`A. Guha, Jpn. J. Genet. 44 (Suppl. 1), 11 (1969).
`31. H. Nijkamp, W. Szybalski, M. Ohashi, W.
`Dove, Mol. Gen. Genet. 114, 80 (1971).
`32. This is paper 2172 from the Laboratory of Ge-
`netics of the University of Wisconsin and paper
`No. 7 in the series "Charon phages for DNA
`Cloning." Paper No. 6 is (2). Supported by NIH
`grpsnt GM21812 to F.R.B., NIH grant CA-07175
`to McArdle Laboratory (W. F. Dove), an NSF
`predoctoral fellowship and an NIH training
`grant to McArdle Laboratory (M.E.F.), NIH
`training grant T32 CA09075 (K.D.-T.), and NIH
`training grant 144-J825 (D.D.M.). We thank W.
`F. Dove for support and advice, Nigel Godsen
`for critical reading of the manuscript, Ed Ko-
`petsky and Brenda Dierschke for technical as-
`sistance, R. Roberts for hospitality in his labora-
`tory, A. Maxam for providing the detailed se-
`quencing procedures and A. Honigman for dis-
`cussions concerning AvaX. This work was done
`under NIH guidelines, which call for EKI, P1.
`19 July 1977; revised 20 September 1977
`
`Expression in Escherichia coli of a Chemically Synthesized
`Gene for the Hormone Somatostatin
`Abstract. A gene for somatostatin, a mammalian peptide (14 amino acid residues)
`hormone, was synthesized by chemical methods. This gene was fused to the Esche-
`richia coli f3-galactosidase gene on the plasmid pBR322. Transformation of E. coli
`with the chimeric plasmid DNA led to the synthesis of a polypeptide including the
`sequence of amino acids corresponding to somatostatin. In vitro, active somatosta-
`tin was specifically cleaved from the large chimeric protein by treatment with cy-
`anogen bromide. This represents thefirst synthesis of a functional polypeptide prod-
`uct from a gene of chemically synthesized origin.
`
`The chemical synthesis of DNA and
`recombinant DNA methods provide the
`technology for the design and synthesis
`of genes that can be fused to plasmid ele-
`ments for expression in Escherichia coli
`or other bacteria. As a model system we
`have designed and synthesized a gene for
`the small polypeptide hormone, soma-
`tostatin (Figs. I and 2). The major con-
`siderations in the choice of this hormone
`were its small size and known amino acid
`sequence (1), sensitive radioimmune and
`biological assays (2), and its intrinsic bio-
`logical interest (3). Somatostatin is a tet-
`radecapeptide; it was originally discov-
`ered in ovine hypothalamic extracts but
`subsequently was also found in signifi-
`cant quantities in other species and other
`tissues (3). Somatostatin inhibits the se-
`cretion of a number of hormones, includ-
`ing growth hormone, insulin, and gluca-
`gon. The effect of somatostatin on the se-
`cretion of these hormones has attracted
`attention to its potential therapeutic val-
`ue in acromegaly, acute pancreatitis, and
`insulin-dependent diabetes.
`The overall construction of the soma-
`tostatin gene and plasmid was designed
`to result in the in vivo synthesis of a pre-
`cursor form of somatostatin (see Fig. 1).
`The precursor protein would not be ex-
`
`pected to have biological activity, but
`could be converted to a functional form
`by cyanogen bromide cleavage (4) after
`cellular extraction. The synthetic soma-
`tostatin gene was fused to the lac operon
`because the controlling sites of this oper-
`on are well characterized.
`Given the amino acid sequence of
`somatostatin, one can design from the
`genetic code a short DNA fragment con-
`taining the information for its 14 amino
`acids (Fig. 2). The degeneracy of the
`code allows for a large number of pos-
`sible sequences that could code for the
`same 14 amino acids. Therefore, the
`choice of codons was somewhat arbi-
`trary except for the following restric-
`tions. First, amino acid codons known to
`be favored in E. coli for expression of the
`MS2 genome were used where appropri-
`ate (5). Second, since the complete se-
`quence would be constructed from a
`number of overlapping fragments, the
`fragments were designed to eliminate un-
`desirable inter- and intramolecular pair-
`ing. And third, G*C-rich (guanine-cyto-
`(adenine-
`sine) followed by A*T-rich
`thymine) sequences were avoided since
`they might terminate transcription (6).
`varying
`oligonucleotides,
`Eight
`in
`length from 11 to 16 nucleotides, labeled
`SCIENCE, VOL. 198
`
`Merck Ex. 1068, pg 1527
`
`

`
`in Fig. 2 as A through H, were synthe-
`sized by the triester method (7). In addi-
`tion to the 14 codons for the structural
`information of somatostatin, several oth-
`er features were built into the nucleotide
`sequence. First, to facilitate insertion in-
`to plasmid DNA, the 5' ends have single-
`stranded cohesive termini for the Eco RI
`and Bam HI restriction endonucleases.
`Second, a methionine codon precedes
`the normal NH2-terminal amino acid of
`somatostatin, and the COOH-terminal
`codon is followed by two nonsense co-
`dons.
`In the cloning and expression of the
`synthetic somatostatin gene we used two
`plasmids. Each plasmid has an Eco RI
`substrate site at a different region of the
`,3-galactosidase
`structural
`(see
`gene
`Figs. 3 and 4). The insertion of the syn-
`thetic somatostatin DNA fragment into
`the Eco RI sites of these plasmids brings
`the expression of the genetic information
`in that fragment under control of the lac
`operon controlling elements. After the
`insertion of the somatostatin fragment
`into these plasmids, translation should
`result in a somatostatin polypeptide pre-
`by
`either
`amino
`acids
`ceded
`ten
`(pSOM 1) or by virtually the whole ,3-ga-
`
`E. coli Lac Operon DNA
`
`GENETIC CODE
`
`Chemical
`DNA Synthesis
`
`(
`
`Lac
`
`P 0
`t~AC
`
`'
`
`j3- Gal
`i
`
`l
`
`lAATTC
`
`Somatostatin Gene
`|ATG I GCT GGT TGT AAG AAC TTC TTTr)
`
`- In In '2 PI PI
`
`PW3(RZZ
`
`Plasmid DNA
`
`MAIA
`
`TAG IGAT AGT I TGT GCT TCA CTT TCA GA
`
`I
`
`| In Vivo
`
`NH2
`
`f-Gal
`
`Meet Ala
`
`Gly
`
`Som
`Cys
`Lys Asn
`s
`
`Phe Phe
`
`Trp
`Lys
`
`Thr
`
`Ser' Thr- Phe
`
`HO Cys
`In Vitro
`Cyanogen Bromide
`Cleavage
`
`f- Gal Fragments
`
`+
`
`NH2 ' Ala
`
`Gly * Cys Lys *Asn Phe Phe.
`gI
`S
`Cys * Ser ' Thr - Phe * Thr-
`HO
`Active Somatostatin
`Fig. 1. Schematic outline of the experimental plan. The gene for somatostatin, made by chem-
`ical DNA synthesis, was fused to the E. coli ,8-galactosidase gene on the plasmid pBR322. After
`transformation into E. coli, the chimeric plasmid directs the synthesis of a chimeric protein that
`can be specifically cleaved in vitro at methionine residues by cyanogen bromide to yield active
`mammalian peptide hormone.
`
`Trp
`Lys
`
`Ala
`
`Gly
`
`Cys
`
`Lys
`
`a
`Barn HI
`Eco RI
`(A)
`(D)
`(B)
`(C)
`BaW H
`~
`*i'
`*
`*00 .4
`4&-
`5A AT T C A T G G C T G G T T G TA A G A A C T T C T T T T G G A A GA C T T T CA C T T C G T G T T GA TAG
`G T A C C GA C C AAC A T T C T T GAAGAAAA C C T T C T GAAAG T GAAGC A C AAC T A T C C T A G '
`(H)
`(G)-
`(F)
`(E)
`' '
`No*
`10 "
`0
`
`Asn
`
`Phe Phe
`
`Trp
`
`Lys
`
`Thr
`
`Phe
`
`Thr
`
`Ser
`
`Cys
`
`Stop
`
`Stop
`
`b
`
`DMT
`
`B
`l
`os4L_.
`o0
`
`[om
`
`1
`
`ci
`
`B
`L
`OR
`
`T
`
`+
`
`HO
`
`B
`
`O
`opIl
`Ps
`
`DMT
`
`\
`
`B
`
`OR
`
`2
`
`3
`
`ci
`
`H+
`
` on June 20, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`Fig. 2. Chemical synthesis of the somatostatin
`gene. (a) Eight oligodeoxyribonucleotides, la-
`beled A through H, were synthesized by the
`modified triester method (7, 23). The codons
`are indicated, and their corresponding amino
`acids are given. The eight fragments were de-
`signed to have at least five nucleotide com-
`plementary overlaps to ensure efficient joining
`by T4 DNA ligase. (b) Recent improvements
`in the synthesis of fully protected trimers,
`which constitute codon blocks and are the
`basic units for building longer oligodeoxyri-
`bonucleotides. With an excess of 1 (2 mmole),
`the coupling reaction with 2 (1 mmole) went
`almost to completion in 60 minutes with the
`aid of a powerful coupling reagent, 2,4,6-
`triisopropylbenzenesulfonyl
`tetrazolide
`(TPSTe, 4 mmole) (2). The 5'-protecting
`group was removed with 2 percent benzene
`sulfonic acid, and the 5'-hydroxyl dimer 5
`could be separated from an excess of 3'-
`phosphodiester monomer 4 by simple solvent
`B = Protected Base
`DMT = 4,4'-dimethoxytrityl
`extraction with aqueous NaHCO3 solution in
`CHC13. The fully protected trimer block was
`prepared successively from the 5'-hydroxyl
`O
`Cl
`11
`11
`dimer 5, 1 (2 mmole), and TPSTe (4 mmole)
`R =-C-<y-OME or -P-O- <-CI
`and isolated by chromatography on silica gel (24). These improvements simplify the purifi-
`I
`cation step and lead to an increase in the overall yields of trimer blocks and to a decrease
`OCH2CH2CN
`in the working time by at least a factor of 2 (21). The eight oligodeoxyribonucleotides then
`were synthesized from the trimers by published procedures (7). The final products, after removal of all protecting groups, were purified by high-
`pressure liquid chromatography on Permaphase AAX (25). The purity of each oligomer was checked by homochromatography on thin-layer
`DEAE-cellulose and also by gel electrophoresis in 20 percent acrylamide (slab) after labeling of the oligomers with [y-32P]ATP in the presence of
`polynucleotide kinase. One major labeled product was obtained from each DNA fragment.
`9 DECEMBER 1977
`
`H +
`
`B
`L
`o'llI
`
`u
`
`1
`
`Y
`
`HO
`
`4
`
`B
`
`B
`O
`HO-T ORmer
`P
`P.-O
`HO
`N0°
`~~~0
`1
`
`Y
`
`5
`
`OR
`
`T
`
`Tnlme
`
`9 DECEMBER 1977
`
`1057~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~15
`1057
`
`Merck Ex. 1068, pg 1528
`
`

`
`0
`
`'-
`0
`
` on June 20, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`'tI~
`
`t
`
`$414C
`I.
`4) CDEo
`
`co
`
`CD
`_aF < o:0
`
`C.D
`
`X4 FI-
`
`4'~
`
`<I-.
`
`4'
`
`V
`
`0
`
`LIDC) U
`
`U CD
`
`A
`
`W U
`
`0
`
`4
`
`0
`
`v
`
`(D .)
`U
`1
`
`V
`
`<I
`
`0 0
`0 0
`
`:t F
`
`<
`
`> FZc~~~U
`
`LI
`
`~~~~CJDU
`(DC.U
`LI
`
`CD
`
`U
`
`<I
`
`lOS8
`
`SCIENCE, VOL. 198
`
`Merck Ex. 1068, pg 1529
`
`

`
` on June 20, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`lactosidase subunit structure (pSOM1 1-3).
`The plasmid
`construction scheme
`(Fig. 3) begins with plasmid pBR322, a
`well-characterized cloning vehicle (8).
`The lac elements were introduced to this
`plasmid by insertion of an Hae III restric-
`tion endonuclease fragment (203 nucle-
`otides) carrying the lac promoter, ca-
`tabolite-gene-activator-protein
`binding
`site, operator, ribosome binding site, and
`the first seven amino codons of the /3-ga-
`lactosidase structural gene (9) (Figs. 3
`and 4). The Hae III fragment was de-
`rived from Xplac5 DNA. The Eco RI-
`cleaved pBR322 plasmid, which had its
`termini repaired with T4 DNA polymer-
`ase and deoxyribonucleotide triphos-
`phates, was blunt-end ligated
`to the
`Hae III fragment to create Eco RI termi-
`ni at the insertion points. Joining of these
`Hae III and repaired Eco RI termini gen-
`erate the Eco RI restriction site'(Figs. 3
`and 4) at each terminus. Transformants
`of E. coli RRl (8) with this DNA were
`selected for resistance to tetracycline
`
`(Tc) and ampicillin (Ap) on 5-bromo4-
`chloro-indolylgalactoside (X-gal) medi-
`um (10). On this indicator medium, colo-
`nies constitutive for the synthesis of 8-
`galactosidase by virtue of the increased
`number of lac operators titrating repres-
`sor, are identified by their blue color.
`Two orientations of the Hae III fragment
`are possible, but these were distin-
`guished by the asymmetric location of an
`Hha restriction site in the fragment. Plas-
`mid pBH 10 was further modified to elim-
`inate the Eco RI endonuclease site distal
`to the lac operator (pBH20).
`The eight chemically synthesized oli-
`godeoxyribonucleotides (Fig. 2) were la-
`beled at the 5' termini with [y-32P]ATP
`(adenosine triphophatase) by T4 polynu-
`cleotide kinase and joined with T4 DNA
`ligase. Through hydrogen bonding be-
`tween the overlapping fragments, the
`somatostatin gene self-assembles and
`eventually polymerizes into larger mole-
`cules because of the cohesive restriction
`site termini. The ligated products were
`
`treated with Eco RI and Barm HI restric-
`tion endonucleases to generate the soma-
`tostatin gene (Fig. 2).
`The synthetic somatostatin gene frag-
`ment with Eco RI and Bam HI termini
`was ligated to the pBH20 plasmid, pre-
`viously treated with the Eco RI and
`Bam HI restriction endonucleases and
`alkaline
`phosphatase. The treatment
`with alkaline phosphatase provides a
`molecular selection for plasmids car-
`rying the inserted fragment (11). Ampi-
`cillin-resistant
`transformants
`obtained
`with this ligated DNA were screened for
`tetracycline sensitivity, and several were
`examined for the insertion of an Eco Rl-
`Bam HI fragment of the appropriate
`size.
`Both strands of the Eco RI-Bam HI
`fragments of plasmids from two clones
`were analyzed by a nucleotide sequence
`analysis (12) starting from the Bam HI
`and Eco RI sites. The sequence analysis
`was extended into the lac-controlling ele-
`ments; the lac fragment sequence was in-
`
`Fig. 3 (facing page, left). Construction of recombinant plasmids. Plas-
`mid pBR322 was used as the parental plasmid (8). Plasmid DNA (S jg)
`was digested with the restriction endonuclease Eco RI. The reaction
`was terminated by extraction with a mixture of phenol and chloro-
`form; the DNA was precipitated with ethanol and resuspended in 50
`)zl of T4 DNA polymerase buffer (26). The reaction-was started by the.
`addition of 2 units of T4 DNA polymerase. The reaction (held for 30
`minutes at 3"C) was terminated by extraction with phenol and chloro-
`form and precipitation with ethanol. The XplacS DNA (3 ,ug) was di-
`gested with the endonuclease Hae III (8). The digested pBR322 DNA
`was blunt-end ligated with the Hae III-digested Xplac5 DNA in a final
`volume of 30 ,lA with T4 DNA ligase (hydroxylapatite fraction) (27) in
`20 mM tris-HCI (pH 7.6), 10 mM MgC92, 10 mM dithiothreitol, and
`0.5 mM ATP for 12 hours at 12°C. The ligated DNA mixture was dia-
`lyzed against 10 mM tris-HCI (pH 7.6) and used to transform E. coli
`strain RRI (8). Transformants were selected for tetracycline resist-
`ance (Tc') and ampicillin resistance (Apr) on antibiotic (20 ,ug/ml)
`minimal X-gal (40 .ug/ml) medium (10). Colonies constitutive for the
`synthesis of 8-galactosidase were identified by their blue color. After
`45 independently isolated blue colonies were screened, three of them
`were found to contain plasmids with two Eco RI sites separated by
`approximately 200 base pairs (28). Plasmid pBH1O was shown to carry
`the fragment in the desired orientation, that is, lac transcription going
`into the Tc' gene of the plasmid. Plasmid pBH1O was further modified
`to eliminate the Eco RI site distal to the lac operator and plasmid
`pBH20 was obtained (29). The nucleotide sequence from the Eco RI
`site into the lac-control region of pBH20 (data not shown), was con-
`firmed. This plasmid was used for cloning the synthetic somatostatin
`gene. Plasmid pBH20 (10 ,ug) was digested with endonucleases
`Eco RI and Bam HI and treated with bacterial alkaline phosphatase
`(0.1 unit of BAPF, Worthington), and incubation was continued for 10
`minutes at 65°C. The reaction mixtures were extracted with a mixture
`of phenol and chloroform, and the DNA was precipitated with ethanol
`(30). Somatostatin DNA (50 pJ of a solution containing 4 Ag/ml) was
`ligated
`with the Bam HI-Eco RI,
`alkaline phosphat4se-treated
`pBH20 DNA in a total volume of 50 ,ul with the use of 4 ufnits of T4
`DNA ligase for 2 hours at 22°C (31). In a control experiment,
`Bam HI-Eco RI alkaline phosphatase-treated pBH20 DNA was ligat-
`ed in the absence of somatostatin DNA under similar conditions. Both
`preparations were used to transform E. coli RRI . Transformants were
`selected on minimal X-gal antibiotic plates. Ten Tce transformants
`were isolated. In the control experiment no transformants were ob-
`tained. Four out of the ten transformants contained plasmids with
`both an Eco RI and a Bam HI site. The size of the small Eco RI-
`Bam HI fragment of these recombinant plasmids was in all four in-
`stances similar to the size of the in vitro prepared somatostatin DNA.
`Base sequence analysis (12) revealed that the plasmid pSOMI had the
`9 DECEMBER 1977
`
`desired somatostatin DNA fragment inserted (data not shown). Be-
`cause of the failure to detect somatostatin activity from cultures car-
`rying plasmid pSOMi, a plasmid was constructed in which the soma-
`tostatin gene could be located at the COOH-terminus of the f-galac-
`tosidase gene, keeping the translation in phase. For the construction
`of such a plasmid, pSOMi (50 ,Ag) was digested with restriction en-
`zymes Eco RI and Pst I. A preparative 5 percent polyacrylamide gel
`was used to separate the large Pst I-Eco RI fragment that carries the
`somatostatin gene from the small fragment carrying the lac control
`elements (12). In a similar way plasmid pBR322 DNA (50 ,ug) was
`digested with Pst I and Eco RI restriction endonucleases, and the two
`resulting DNA fragments were purified by preparative electrophoresis
`on a 5 percent polyacrylamide gel. The small iPst I-Eco RI fragment
`from pBR322 (I jug) was' ligated with the large Pst I-Eco RI DNA
`fragment (5 Ag) from pSOMI. The ligated mixture was used to trans-
`form E. coli RRI, and transformants were selected for Apr on X-gal
`medium. Almost all the Apr transformants (95 percent) gave white
`colonies (no lac operator) on X-gal indicator plates. The resulting
`plasmid, pSOM I I, was used in the construction of plasmid pSOM I 1-
`3. A mixture of 5 ,ug of pSOMl I DNA and 5 ,.&g of XplacS DNA was
`digested with Eco RI. The DNA was extracted with a mixture of phe-
`nol and chloroform; the extract was precipitated by ethanol, and the
`precipitate was resuspended in T4 DNA ligase buffer (50 ul) in the
`presence of T4 DNA ligase (1 unit). The ligated mixture was used to
`transform E. ,coli strain RR . Transformants were selected for Apr on
`X-gal plates containing ampicillin and screened for constitutive /3-ga-
`lactosidase production. Approximately 2 percent of the colonies were
`blue (such as pSOM1-1 and 11-2). Restriction enzyme analysis of
`plasmid DNA obtained from these clones revealed that all the plas-
`mids carried a new Eco RI fragment of approximately 4.4 megadal-
`tons, which carries the lac operon control sites and most of the 38-
`galactosidase gene (13, 14). Two orientations of the Eco RI fragment
`are possible, and the asymmetric location of a Hind III restriction in
`this fragment can indicate which plasmids had transcription pro-
`ceeding into the somatostatin gene. The clones carrying plasmids
`pSOM I 1-3, pSOM I 1-5, pSOM I 1-6, and pSOM I 1-7 contained the
`Eco RI fragment in this orientation.
`
`Fig. 4 (facing page, right). Nucleotide sequences of the lac-somatosta-
`tin plasmids. The nucleotide sequence of the lac control elements, 38-
`galactosidase structural gene, and the synthetically derived soma-
`tostatin DNA, are depicted (9, 14, 27) along with the restriction endo-
`nuclease substrate sites. The nucleotide sequence of pSOM I, as de-
`picted, was confirmed (legends to Figs. 3 and 5). The nucleotideo se-
`quence of pSOMI 1-3 was inferred from published data (9, 13, 14, 27).
`The amino acid sequence of somatostatin is italicized. The amino acid
`sequence numbers of /3-galactosidase are in brackets.
`
`1059
`
`Merck Ex. 1068, pg 1530
`
`

`
` on June 20, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`than 2 ,ug of protein from formic acid-
`bacterial
`treated
`interfered
`extracts
`somewhat by increasing the background,
`but cyanogen bromide cleavage greatly
`reduced this interference. Reconstruc-
`tion experiments showed that somatosta-
`tin is stable in cyanogen bromide-treated
`extracts.
`analysis
`The DNA sequence
`of
`pSOM indicated that the clone carrying
`this plasmid should produce a peptide
`containing somatostatin. However, to
`date all attempts to detect somatostatin
`radioimmune activity from extracts of
`cell pellets or culture supernatants have
`been
`Negative
`unsuccessful.
`results
`were also obtained when the growing
`culture was added directly to 70 percent
`formic acid and cyanogen bromide. We
`calculate that E. coli RRI (pSOMI) con-
`tains less than six molecules of soma-
`tostatin per cell. In a reconstruction ex-
`periment we have observed that exoge-
`is degraded very
`nous somatostatin
`rapidly by E. coli RR1 extracts. The fail-
`ure to find somatostatin activity might be
`accounted for by intracellular degrada-
`tion by endogenous proteolytic
`en-
`zymes.
`If the failure to detect somatostatin ac-
`tivity from pSOMI was due to pro-
`teolytic degradation of the small protein
`(Fig. 4), attachment to a large protein
`might stabilize it. The f3-galactosidase
`structural gene has an Eco RI site near
`the COOH-terminus (13). The available
`data on the amino acid sequence of this
`protein (13, 14) suggested that it would
`be possible to insert the Eco RI-Bam HI
`somatostatin gene into the site and main-
`tain the proper reading frame for the cor-
`rect translation of the somatostatin gene
`(Fig. 4).
`The construction of this plasmid is
`outlined in Fig. 3. The Eco RI-Pst frag-
`ment of the pSOMi plasmid, with the
`lac-controlling element, was removed
`and replaced with the Eco RI-Pst frag-
`ment of pBR322 to produce the plas-
`mid pSOM I. The Eco RI fragment of
`XplacS, carrying the lac operon control
`region and most of the f8-galactosidase
`structural gene, was inserted into the
`Eco RI site of pSOM I. Two orienta-
`tions of the Eco RI lac fragment of
`XplacS were expected. One of these ori-
`entations would maintain the proper
`reading frame into
`the
`somatostatin
`gene, the other would not.
`A number of independently isolated
`clones
`(with
`designations
`plasmid
`pSOM1 1-2 and pSOMI 1-3) were ana-
`lyzed for somatostatin activity, as de-
`scribed above. In constrast to the results
`of experiments with pSOM1, four clones
`(pSOM11-3, 11-5, 11-6, and 11-7) were
`SCIENCE, VOL. 198
`
`tact, and in one case, pSOMI, the nucle-
`otide sequence of both strands were in-
`dependently determined, each giving the
`sequence shown in Fig. 3. In the other
`case, the sequence was identical except
`for a base pair deletion (A*T) at a posi-
`tion equivalent to the junction of the B-C
`oligonucleotides in the original DNA
`fragment. The basis for the deletion is
`unclear.
`The standard radioimmune assays
`(RIA) for somatostatin (2) were modified
`by decreasing the assay volume and by
`using phosphate buffer (Fig.
`6). This
`
`modification proved suitable for the de-
`tection of somatostatin in E. coli ex-
`tracts. Bacterial cell pellets, extracts, or
`cultures were treated overnight in 70 per-
`cent formic acid containing cyanogen
`bromide (5 mg/ml). Formic acid and cy-
`anogen bromide were removed under
`vacuum over KOH before the assay. Ini-
`tial experiments with extracts of E. coli
`strain RRI (the recipient strain) (10) in-
`dicated that less than 10 pg of somatosta-
`tin could easily be detected in the pres-
`ence of 16 ,tg or more of cyanogen bro-
`mide-treated bacterial
`protein. More
`
`Fig. 5. Ligation and acrylamide gel analysis of somatostatin DNA. The 5'-OH termini of the
`chemically synthesized fragments A through H (Fig. 2a) were labeled and phosphorylated sepa-
`rately. Just prior tu the kinase reaction, 25 ,c of [y-32P]ATP (- 1500 c/mmole) (12) was evapo-
`rated to dryness in 0.5-ml Eppendorf tubes. The fragment (5 Ag) was incubated with 2 units of
`T4 DNA kinase (hydroxylapatite fraction, 2500 unit/ml) (26), in 70 mM tris-HCI, pH 7.6, 10 mM
`MgCl2, and 5 mM dithiothreitol in a total volume of 150 ,ul for 20 minutes at 37°C. To ensure
`maximum phosphorylation of the fragments for ligation purposes, 10 ,lI of a mixture consisting
`of 70 mM tris-HCI, pH 7.6, 10 mM MgC92, 5 mM dithiothreitol, 0.5 mM ATP, and 2 units of
`DNA kinase were added, and incubation continued for an additional 20 minutes at 37°C. The
`fragments (250 ng/Al) were stored at -20°C without further treatment. Kinase-treated fragments
`A, B, E, and F (1.25 ,ug each) were ligated in a total volume of 50 Al in 20 mM tris-HCI (pH 7.6),
`10 mM MgCl2, 10 mM dithiothreitol, 0.5 mM ATP, and 2 units of T4 DNA ligase (hydroxyl-
`apatite fraction, 400 unit/ml) (26), for 16 hours at 4°C. Fragments C, D, G, and H were ligated
`under similar conditions. Samples (2 ,A) were removed for analysis by electrophoresis on a 10
`percent polyacrylamide gel and subsequent autoradiography (16) (lanes 1 and 2, respectively).
`The fast migrating material represents unreacted DNA fragments. Material migrating with the
`bromophenol blue dye (BPB) is the monomeric form of the ligated fragments. The slowest
`migrating material represents dimers, which form by virtue of the cohesive ends, of the ligated
`fragments A, B, E, and F (lane 1) and C, D, G, and H (lane 2). The dimers can be cleaved by
`restriction endonuclease Eco RI or Bam HI, respectively (data not shown). The two half mole-
`cules (ligated A + B + E + F and ligated C + D + G + H) were joined by an additional liga-
`tion step carried out in a final volume of 150 IlI at 4°C for 16 hours. A sample (1 JAI) was removed
`for analysis (lane 3). The reaction mixture was heated for 15 minutes at 65°C to inactivate the T4
`DNA ligase. The heat treatment does not affect the migration pattern of the DNA mixture (lane
`4). Enough restriction endonuclease Bam HI was added to the reaction mixture to cleave the
`multimeric forms of the somatostatin DNA in 30 minutes at 37°C (lane 5). After the addition of
`NaCl to a concentration of 100 mM, the DNA was digested with Eco RI endonuclease (lane 6).
`The restriction endonuclease digestions were terminated by phenol-chloroform extraction of
`the DNA. The somatostatin DNA fragment was purified from unreacted and partially ligated
`DNA fragments by preparative electrophoresis on a 10 percent polyacrylamide gel. The band
`indicated with an arrow (lane 7) was excised from the gel, and the DNA was eluted by slicing the
`gel into small pieces and extracting the DNA with elution buffer (0.SM ammonium acetate,
`10 mM MgCI2, 0.1 mM EDTA, and 0.1 percent sodium dodecyl sulfate) overnight at 65°C (12).
`The DNA was precipitated with two volumes of ethanol, centrifuged, redissolved in 200 plI of
`10 mM tris-HCI (pH 7.6), and dialyzed against the same buffer, resulting in a somatostatin DNA
`concentration of 4 ,ug/ml.
`1060
`
`Merck Ex. 1068, pg 1531
`
`

`
` on June 20, 2016
`
`http://science.sciencemag.org/
`
`Downloaded from
`
`entation of the lac operon and the pro-
`duction of somatostatin radioimmune
`activity.
`The design of the somatostatin plas-
`mid predicts that the synthesis of soma-
`tostatin would be under the control of
`the lac operon. The lac repressor gene is
`not included in the plasmid, and the re-
`cipient strain (E. coli RR1) contains the
`wild-type chromosomal lac repressor
`gene, which produces only 10 to 20 re-
`pressor molecules per cell (15). The plas-
`mid copy number (and therefore the
`number of lac operators) is approximate-
`ly 20 to 30 per cell and complete repres-
`sion is impossible. The specific activity
`of somatostatin in E. coli RR1 (pSOMl 1-
`3) was increased by IPTG, an inducer of
`the lac operon (Table I). As expected,
`the level of induction was low, varying
`from 2.4- to 7-fold. In experiment 7
`(Table 1), the a activity (14), a measure
`of the first 92 amino acids of f-galactosi-
`dase, also was induced by a factor of 2.
`
`In several experiments (Table 1 and oth-
`er experi