Contributors to V o l u m e 217 Article numbers are in parentheses following the names of contributors. Affiliations li...
33 downloads
1398 Views
21MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Contributors to V o l u m e 217 Article numbers are in parentheses following the names of contributors. Affiliations listed are current.
GIOVANNA F E R R O - L u z z 1 A M E S ( 3 2 ) , Department of Molecular and Cell Biology, Division of Biochemistry, University of California, Berkeley, Berkeley, California 94720
lina, Chapel Hill, North Carolina 27599 RICHARD L. CATE (29), Biogen, Inc., Cambridge, Massachusetts 02142 KARL X. CHA1 (23), Department qf Biochemistry and Molecular Biology, Medical University of South Carolina. Charleston, South Carolina 29425 JULIE CHAO (23), Department of Biochemistry and Molecular Biology, Medical University of South Carolina, Charleston, South Carolina 29425
SILV1A B~HRING (5), Institutfi~r Molekularhiologie, Abteilung Molekulare Zellgenetik, D-Ill5 Berlin-Buch, Germany VLADIMIR [. BARANOV (9), RiboGene, Inc., Hayward, California 94545 CARL A. BATT (18), Department of Food Science, Cornell University, Ithaca, New York, 14853
LEE CHAO (23), Department of Biochemistry and Molecular Biology, Medical University of South Carolina, Charleston. South Carolina 29425
JEAN-PAUL BEHR (41), Laboratoirede Chi. mie Gdndtique, Universitd Louis Pasteur, CNRS URA 1386, F-67401 lllkirch, France
LIN CHEN (7), Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138
MARTIN W. BERCHTOLD (8), lnstitut fiir Pharmakologie und Biochemie, Universitiit Ziirich-lrchel, Ch-8057 Zurich, Switzerland
YUNJE CHO (18), Field of Microbiology. Cornell University, Ithaca, New York 14853 CHRISTOPHER COLECLOUGH (11), Department of Immunology, St. Jude Children's Research Hospital, Memphis, Tennessee 38101 MATTHEW COTTEN (42), Research Institute of Molecular Pathology, A-I030 Vienna. Austria
MAX L. BIRNSTIEL (42), Research Institute of Molecular Pathology, A-I030 Vienna, Austria JOHN E. BOYNTON (37), Department of Botany, Duke University, Durham, North Carolina 27706 |RENA BRONSTEIN (29), Tropix, Inc., Bedford, Massachusetts 01730 LAKI BULUWELA (28), Department of BiDchemistry, Charing Cross and Westminster Medical School, London W6 8RF, England ZELING CAI (17), Department oflmmunology, Mayo Clinic, Rochester, Minnesota 55905 CELESTE CANTRELL (31), Department of Pharmacology, University of North Caroix
RICHARD G. H. COTTON (19), Olive Miller Protein Laborato~, Murdoch Institute. Royal Children's Hospital, Parkville Victoria 3052. Australia HENRY DANIELL (38), Department of Botany and Microbiology, Auburn Universit3", Auburn, Alabama 36849 BIMALENDU DASMAHAPATRA (10), Department of Antiviral Chemotherapy, Schering-Plough Research Corporation, Bloomfield, New Jersey 07003
X
CONTRIBUTORS TO VOLUME 217
NORMAN DAVIDSON (33), Division of Biol-
MICKEY C-T. Hu (33), Department of Ex-
ogy, California Institute of Technology, Pasadena, California 91125 ANTONIA DESTREE (39), Therion Biologics Corporation, Cambridge, Massachusetts 02142 V. J. DWARK! (43), Vical Inc., San Diego, California 92121 FRITZ ECKSTEIN (13), Abteilung Chemie, Max-Planck-lnstitut fiir Experimentelle Medizin, D-3400 GOttingen, Germany CHRISTIAN W. EHRENFELS (29), Biogen, Inc., Cambridge, Massachusetts 02142 J. VICTOR GARCIA (40), Department of Virology and Molecular Biology, St. Jude Children's Research Hospital, Memphis, Tennessee 38101 NICHOLAS W. GILLHAM (37), Department of Zoology, Duke University, Durham, North Carolina 27706 ALEXANDER N. GLAZER (30), Department of Molecular and Cell Biology, Division of Biochemistry and Molecular Biology, University of California, Berkeley, Berkeley, California 94720 MICHAEL M. GOTTESMAN (4), Laboratory of Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892 RICHARD P. HAUGLAND (30), Molecular Probes, Inc., Eugene, Oregon 97402 STEFEAN N. Ho (17), Department of Pathology, Stanford University Medical School, Stanford, California 94305 BERND HOFER (12), Abteilung Mikrobiologie, Gesellschaft far Biotechnologische Forschung, D-3300 Braunschweig, Germany CHRISTA HORICKE-GRANDPIERRE (6), Abteilung Genetische Grundlagen der Pflanzenziichtung, Max-Planck-lnstitut fiir Ziichtungsforschung, D-5000 KOIn 30, Germany
perimental Hematology, Amgen, Inc., Amgen Center, Thousand Oaks, California 91320 TIM C. HUFFArER (21), Section of Biochemistry, Molecular and Cell Biology, Cornell University, Ithaca, New York 14853 HENRY D. HUNT (17), Department of Immunology, Mayo Clinic, Rochester, Minnesota 55905 ANDREW C. JAMIESON (18), Melvin Calvin Laboratory, University of California, Berkeley, California 94730 R. JILK (22), Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, Wisconsin 53706 SUSAN E. KANE (4), City of Hope National Medical Center, Duarte, California 91010 PETR KARLOVSKY (24), Institute of Plant Pathology, University of GOttingen, D-3400 Gdttingen, Germany DAVID C. KASLOW (20), Molecular Vaccine Section, Laboratory of Malaria Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892 M. P. KREBS (22), Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
ROBERT M. HORTON (17), Department of
Biochemistry, Gortner Laboratories, University of Minnesota, St. Paul, Minnesota 55108
BIRGIT Kt)HLEIN (12), Max-Planck-lnstitut
far Experimentelle Endocrinologie, D-3000 Hannover, Germany ERIC LAI (31), Department of Pharmacol-
ogy, University of North Carolina School of Medicine, Chapel Hill, North Carolina 27599 ANDRE LIEBER (5), Abteilung Molekulare
Zellgenetik, lnstitut fiir Molekularbiologie, D-1115 Berlin-Buch, Germany JEAN-PHILIPPE LOEFFLER (41), lnstitut de
Physiologie, CNRS URA 1446, F-67084 Strasbourg, France
CONTRIBUTORS TO VOLUME 217
Xi
CARMEL M. LYNCH (40), Targeted Genetics
HENRig 0RUM (2), Department of Biochem-
Corporation, Seattle, Washington 98101 CHRISTOPH MAAS (6), Abteilung Genetische Grundlagen der Pflanzenziichtung, MaxPlanck-lnstitut fiir Ziichtungsforschung, D-5000 KOln 30, Germany KURTIS D. MACFERRIN (7), Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138 KAYO MAEDA (1), European Molecular Biology Laboratory, Hamburg Outstation, D-2000 Hamburg, Germany ANNA MASR (39), Integrated Genetics, Inc., Framingham, Massachusetts 01701 J. C. MAKRIS* (22), Lawrence Livermore National Laboratory, Livermore, California, 94551 ROBERT W. MALONE (43), Department of Pathology, University of California, Davis Medical Center, Sacramento, Califi~rnia, 95817 RICHARD A. MATHIES (30), Department of Chemistry, University of California, Berkeley, Berkeley, California 94720 GAIL P. MAZZARA (39), Therion Biologics Corporation, Cambridge, Massachusetts 02142 A. DUSTY MILLER (40), Program in Molecular Medicine, The Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 DANIEL G. MILLER (40), Program in Molecular Medicine, The Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 CESAR MILSTEIN (28), Medical Research Council Laboratory of Molecular Biology, Cambridge CB2 2QH, England OWEN J. MURPHY (29), Tropix, Inc., Bedford, Massachusetts 01730 P. L. NORDMANN (22), Department of Microbiology, Biozentrum, University of Basel, CH-4056 Basel, Switzerland DAVID B. OLSEN (13), Merck Sharp and Dohme, Research Laboratories, West Point, Pennsylvania 19486
istry B, The Panum Institute, Research Center for Medical Biotechnology, University of Copenhagen, DK-2200 Copenhagen N, Denmark
* Deceased.
GARY V. PADDOCK (25), Department of Mi-
crobiology and Immunology, Medical University of South Carolina, Charleston, South Carolina 29425 R. PADMANABHAN (14), Department of Bio-
chemistry and Molecular Biology, University of Kansas Medical Center. Kansas City, Kansas 66013 THOMAS L. PAULS (8), lnstitutfiir Pharma-
kologie und Biochemie, Universitdt Zt~rich-lrchel, CH-8057 Zurich, Switzerland LARRY R. PEASE (17), Department of Immu-
nology, Mayo Clinic', Rochester, Minnesota 55905 HUNTINGTON POTTER (34), Department of
Neurobiology, Harvard Medical School, Boston, Massachusetts 02115 LAgs K. POULSEN (2), Department of Mi-
crobiology, Denmark Technical University, DK-2800 Lyngby, Denmark ANNEMARIE POUSTKA (26, 27), lnstitut j'~r
Virusforschung, Angewandte Tumorvirologie, Deutsches Krebsforschungszen(rum, D-6900 Heidelberg, Germany JEFFREY K. PULLEN (17), Department of
Immunology, Mayo Clinic, Rochester. Minnesota 55905 MARK A. QUESADA (30), Department of
Chemistry, University of California, Berkeley, Berkeley, California 94720 DAVID J. RAWLINGS (20), Howard Hughes
Medical Institute, University of California, Los Angeles, Los Angeles, California 90024 W. S. REZNIKOFF (22), Department of Bio-
chemistry, College of Agricultural and Life Sciences, University of WisconsinMadison, Madison, Wisconsin 53706 J. A. RUSSELL (36), Department of Horticul-
tural Sciences, New York State Agricul-
xii
CONTRIBUTORS TO VOLUME 217
tural Experiment Station, Cornell University, Geneva, New York 14456 HAYS S. RYE (30), Department of Molecular
and Cell Biology, Division of Biochemistry and Molecular Biology, University of California, Berkeley, Berkeley, California 94720 JENNIFER A. SALEEBA (19), Department of Biological Science, Dartmouth College, Hanover, New Hampshire 03755 VOLKER SANDIG (5), lnstitutfiir Molekularbiologie, Abteilung Molekulare Zellgenetik, D-1115 Berlin-Buch, Germany J. C. SANFORD (36), Department of Horti-
cultural Sciences, New York State Agricultural Experiment Station, Cornell University, Geneva, New York 14456 JON R. SAYERS (13), School of Biological Science, University of North Wales, Bangor, Gwynedd, Wales LL57 2DG JEFF SCHELL (6), Abteilung Genetische
Grundlagen der Pflanzenziichtung, MaxPlanck-lnstitut fiir Ziichtungsforschung, D-5000 KOln 30, Germany STUART L. SCr~REIRER (7), Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138 JAMIE K. SCOTT (15), Division of Biological
Sciences, University of Missouri, Columbia, Missouri 65211 GEORG SCZAKIEL (1), Angewandte Tumorvirologie, Deutsches Krebsforschungszentrum, D-6900 Heidelberg, Germany VENKATAKRISHNASHYAMALA(32), Chiron
Corporation, Emeryville, California 94608 JOHN R. SIMON (35), Department of Biologi-
cal Chemistry and Laboratory of Biomedical & Environmental Sciences, University of California School of Medicine, Los Angeles, California 90024 F. D. SMITH (36), Department of Horticul-
tural Sciences, New York State Agricultural Experiment Station, Cornell University, Geneva, New York 14456 GEORGE P. SMITH (15), Division of Biologi-
cal Sciences, University of Missouri, Columbia, Missouri 65211
WOLFGANG SOMMER (5), lnstitut far Mole-
kularbiologie, Abteilung Molekulare Zellgenetik, D-Ill5 Berlin-Buch, Germany ALEXANDER S. SPIRIN (9), Institute of Protein Research, Academy of Sciences, 142292 Pushchino, Moscow Region, Russia HANS-HENNING STEINBISS (6), Abteilung Genetische Grundlagen der Pflanzenziichtung, Max-Planck-lnstitut far Ziichtungsforschung, D-5000 KOln 30, Germany MICHAEL STRAUSS (5), Max-Planck Group of the Humboldt University, MaxDelbriick Center for Molecular Medicine, D-I 115 Berlin-Buch, Germany MICHAEL P. TERRANOVA(7), Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138 RICHARD TIZARD (29), Biogen, Inc., Cambridge, Massachusetts 02142 REINHARD TOPFER (6), Abteilung Genetische Grundlagen der Pflanzenziichtung, Max-Planck-lnstitut far Ziichtungsforschung, D-5000 KOln 30, Germany SHIGEZO Ut)AKA (3), Department of Food Science and Technology, Faculty of Agriculture, Nagoya University, Nagoya 464, Japan GREGORY L. VERDINE (7), Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138 INDER M. VERMA (43), Molecular Biology and Virology Laboratory, The Salk Institute, San Diego, California 92186 JOHN C. VOYTA (29), Tropix, Inc., Bedford, Massachusetts O1730 ERNST WAGNER (42), Research Institute of Molecular Pathology, A-I030 Vienna, Austria MARY M. Y. WAYE (16), Department of Biochemistry, Chinese University of Hong Kong, Hong Kong M. WEINREICH (22), Department of Biochemistry, College of Agricultural and Life Sciences, University of WisconsinMadison, Madison, Wisconsin 53706
CONTRIBUTORS TO VOLUME 217 T. WIEGAND (22), Department of Biochem-
istry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, Wisconsin 53706 LAI-CHu W c (28), Davis Medical Center, Departments of Medical Biochemistry and Internal Medicine, The Ohio State University, Columbus, Ohio 43210 HIDEO YAMAGATA (3), Department of Food Science and Technology, Faculty of Agriculture, Nagoya University, Nagoya 464, Japan C. YUNG YU (28), Departments of Pediatrics and Medical Microbiology and lm-
Xlll
munology, The Ohio State University and Children's Hospital Research Foundation, Columbus, Ohio 43205 STEPHEN YUE (30), Molecular Probes, Inc.,
Eugene, Oregon 97402 Q.-X. ZHANG (14), Department of Biochem-
istry and Molecular Biology, University of Kansas Medical Center, Kansas City, Kansas 66103 L.-J. ZHAO (14), Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, Kansas 66103
[1]
E. coli EXPRESSIONPLASMIDpPLEX
[1] V e c t o r p P L E X
By
3
for E x p r e s s i o n o f N o n f u s i o n P o l y p e p t i d e s in E s c h e r i c h i a coli
GEORG SCZAKIEL a n d KAYO MAEDA
Introduction Escherichia coli bacteria are a powerful tool for the production of heterologous proteins in large quantities, which is of general experimental importance in many fields of natural sciences, for example, in biochemical and biophysical studies. The functional genes coding for polypeptides of interest are introduced stably into E. coli bacteria by E. coli vectors (e.g., plasmids, bacteriophages, cosmids, and phagemids). The expressed polypeptides originate from a unique type of coding DNA and thus, in E. coli from nonspliceable mRNAs, the peptide sequence of expressed molecules is defined exactly, that is, they are monoclonal. For many studies, monoclonal polypeptides are of great advantage in comparison with protein preparations from natural sources, which may consist of numerous closely related but not identical isoforms. Escherichia coli is one of the best studied organisms and many well-established methodologies used in molecular biology can be applied to modify and handle vectors and coding sequences. 1,2 Polypeptides of interest can be expressed in E. coli as fusion proteins, usually extended at the amino terminus with prokaryotic portions intended to provide increased translational initiation, stability, solubility, alternative purification protocols, and yield, or to allow secretion. Fusion proteins can be used for immunological studies, such as the production of antisera, or as antigens in enzyme-linked immunosorbent assay (ELISA) or Western analysis. However, their use in other studies, for example, those concerning enzymatic activities and three-dimensional structures, is limited, especially in the latter case, where the expression of nonfusion proteins is desirable. The necessary elements that an expression plasmid should supply are an origin of replication, a dominant selection marker for plasmid propagation and maintenance, and transcriptional (promoter) and translaLT. Maniatis, E. F. Fritsch, and J. Sambrook, "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982. 2 F. M. Ausubel, R. Breut, R. E. Kingston, D, D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl, "Current Protocols in Molecular Biology." Wiley, New York, 1987.
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
4
VECTORS FOR EXPRESSING CLONED GENES
[1]
tional initiation sites (Shine-Dalgarno sequence and start codon), as well as termination signals for translation and transcription. Transcription directed by strong promoters can down-regulate plasmid replication, which may result in the loss ofplasmid. For this reason transcription from strong promoters usually needs to be terminated by efficient transcriptional terminators, A number of other parameters for successful expression of heterologous eukaryotic sequences in E. coli must be considered and tested: (1) DNA sequence and primary and secondary structure of the transcript in the vicinity of the start codon, 3 (2) codon usage, 4 (3) possible toxicity of expression products for E. coli, (4) posttranslational modifications, (5) RNA editing of eukaryotic sequences in the homologous system, 5'6 which does not occur in E. coli, and (6) evaluation of the ability of expressed portions of proteins to form defined structures. The techniques for prokaryotic gene expression have been described in detail. 7 Principle of Method The expression vector pPLEX 8 contains all elements necessary for the expression of open reading frames in E. coli. For transcription the bacteriophage h-derived strong PL Promoter9 and the t R terminator are used. The PL promoter can be regulated, that is, repressed or induced by the thermolabile h ci857 repressor,10 which is active at the permissive temperature of 28 ° but is inactive at 37 or 42 °. The gene coding for the ci857 repre s sor can be plasmid encoded or can be integrated into the host cell chromosome (e.g., E. coli strain NF 1). The translational control elements, that is, the ribosomal binding site and stop codons in all three reading frames as well as unique cloning sites in between, are indicated in Fig. 1. Materials and Methods
Escherichia coli Strains NF1 (K12) AH1 H): F - A(bio- uvrB) lacZam hNam7 Nam53 ci857 AH1 (cro-F-A-J-b2 ) 3 H. A. De Boer and A. S. Hui, this series, Vol. 185, p. 103. 4 p. M. Sharp and W.-H. Li, Nucleic Acids Res. 15, 1281 (1987). 5 L. Simpson and J. Shaw, Cell 57, 355 (1989). 6 A. M. Weiner and N. Maizels, Cell 61, 917 (1990). 7 D. V. Goeddel, this series, Vol. 185, p. 3. 8 G. Sczakiel, A. Wittinghofer, and J. Tucker, Nucleic" Acids Res. 15, 1878 (1987). 9 E. Remaut, P. Stanssens, and W. Fiers, Gene 15, 81 (1981). l0 M. Lieb, J. Mol. Biol. 16, 149 (1966). ii H.-U. Bernard, E. Remaut, M. V. Hershfield, H. K. Das, D. R. Helinski, C. Yanowsky, and N. Franklin, Gene 5, 59 (1979).
[1]
E. coli EXPRESSION PLASMID p P L E X
5
(130) NCOI
SalI
HindIII
HpaI
BclI
I CCATr~GTCGAC AAG CTT AC;TTAACTOATCA (o)
~
Stul
Pvu[
/
/
/ I-
\
/
"\ \
(3450)"~ ( {
PstI l \
/
pPLEX
i
\
/ /¢~/
fO /
/
(~ Ms 2 Shine- Datgarno Sequence: EcoR 1 GAATTCCGAC
TGCGAGCTTA
TTGTTAAGGC
AATGCAAGGT
CTCCTAAAAG
ATGGAAACCC
GATTCCCTCA
GCAATCGCAG
CAAACTCCGG
CATCTACTAA
TAGACGCCGG
CCATTCAAAC
ATGAGGATTA
CCCATGG
Nco 1 ®
%tR
Sequence:
TAAATAACCC
CGCTCTTACA
CATTCCAGCC
CTGAAAAAGG
Nsi I GCATCAAATT
AAACCACACC
TATGGTGTAT
GCATACATTC
AATCAATTGT
TATCTAAGGA A A T A C T T A C A
GCATTTATTT TATG
FIG. 1. Structure of the E. coli expression plasmid pPLEX and sources of sequence elements: A, MS 2 Shine-Dalgarno (S.D.) sequence [G. Simons, E. Remaut, B. Allet, R. Devos, and W. Fiers, Gene 28, 55 (1984)]; B, htR fragment; C, galactokinase gene [C. Debouck, A. Riccio, D. Schlumperli, K. McKenney, J. Jeffers, C. Hughes, and M. Rosenberg, Nuclei(" Acids Res. 13, 1841 (1985)]; D, fragment from pPLc245 containing the ,kpL promoter [E. Remaut, P. Stanssens, and W. Fiers, Nucleic Acids Res. 11, 4677 (1983)]. Note that BclI is sensitive to Dam methylation. In order to use the BclI site pPLEX must be grown in a dam- E. coli strain. An additional AccI site located on the pBR322 sequence that is present in the original plasmid pPLEX but was filled in with Klenow fragment and nucleotide triphosphates, that is, it was destroyed in pPLEXAcc • (J. Tucker, unpublished observations, 1986.)
6
VECTORS FOR EXPRESSING CLONED GENES
[1]
W6 (origin not known): su-, cI (wild type) unc195912: lacI Q lacL8 thr-1 ara-14 leuB6 A(gpt-proA) 62 lacY1 1on-22 supE44 galK2 h- sulA27 hisG4 rpsL31 xyl-5 mtl-1 thi-1
Cloning Methods of recombinant DNA technology are essentially performed following the protocols of Maniatis et al.1 For cloning pPLEX-derived constructs we use E. coli strain W6, containing the h wild-type cI repressor integrated into its chromosome. The wild-type repressor is able to shut off the PL promoter efficiently, thus allowing stable replication and high copy numbers of recombinant pPLEX-derived constructs. In principle an E. coli strain harboring the thermolabile ci857 repressor is also suitable at the permissive temperature of 28°; however, the clearly decreased growth rate at this temperature seems to be a disadvantage. For induction of the PL promoter, E. coli host strains NFI and unc1959, both containing a cI857-carrying plasmid, are used. Transformation of E. coli cells is performed following the CaCI2 method 13 for W6 and NF1 or the protocol developed by Hanahan 14 for DH2/6. The transformation yields for 1 /xg of pPLEX DNA with freshly prepared bacteria are in the range of 5 × 105 for W6 and 1 × 106 for NF1. The transformation frequency after storage of transformation-competent cells in 5% (v/v) glycerol at - 7 0 ° is decreased by a factor of approximately 10.
Induction of hPL Promoter The protocol for the induction of the hPL promoter of E. coli strains carrying pPLEX constructs is depicted schematically in Fig. 2. As an alternative way of induction of the LMM expression plasmid pEXLMM74 a temperature shift to 42 ° may be performed for 15 min with subsequent incubation at 37° for 4 hr. To raise the temperature quickly to 42 ° for large volumes (e.g., 10 liter), an appropriate amount of fresh medium preheated to 60 ° is added. On induction, suppression of the htR terminator results in transcription of a bicistronic mRNA consisting of the heterologous open reading frame and the coding sequence for galactokinase. Thus, an increase in galactokinase activity monitors efficient hpL-directed transcription.
12 Obtained from B. Bachman, E. coli Genetic Stock Centre, New Haven, Connecticut. I3 M. Mandel and A. Higa, J. Mol. Biol. 53~ 159 (1970). 14 D. Hanahan, J. Mol. Biol. 166, 557 (1983).
[1]
E. coli EXPRESSION PLASMID pPLEX
7
Grow 1 ml overnight culture of E. coli strain NF1 transformed with pPLEX construct in medium (standard I or L-broth supplemented with 100 p.g/ml ampiciUin) at 28° $ Inoculate 1 ml of fresh medium with 10/xl of dense overnight culture Incubate for 1 hr at 28° $ Divide culture in two 0.5-ml aliquots
/
\
4 hr, 28° (uninduced control)
4 hr, 28° (induced control)
l
Protein analysis
Protein analysis
1
FIG. 2. Protocol for the induction of the PL promoter-driven expression cassette of pPLEX. In analysis of expression products by SDS-polyacrylamide gel electrophoresis induced cultures have higher cell densities, i.e., protein concentrations, than do control cultures grown at 28°.
Analysis o f Expression Products Soluble Protein Fraction. Escherichia coti cells are harvested by centrifugation (30 sec, r o o m t e m p e r a t u r e , 7000 rpm, E p p e n d o r f centrifuge) and the cell pellet is r e s u s p e n d e d with 1 ml 50 m M Tris-HCl (pH 7.4). After centrifugation the pellet is r e s u s p e n d e d vigorously in 0.5 ml lysis buffer containing 50 m M Tris-HCl (pH 7.4), 0.5 m M dithioerythritol (DTE), 0. I m M phenylmethylsulfonyl fluoride (PMSF), and 1 m M ethylenediaminetetraacetic acid (EDTA). L y s o z y m e (3/zl, 10 mg/ml in 10 m M Tris-HCl p H 8.0, 1 m M E D T A ) is added and the mixture is maintained for 10 to 20 min at r o o m t e m p e r a t u r e . Sodium d e o x y c h o l a t e (3/zl, 40 mg/ml in water) is added and the solution is kept for 15 min at r o o m temperature. After centrifugation (15 min, 4 °, E p p e n d o r f centrifuge) soluble proteins are contained in the clear supernatant. Sodium Dodecyl Sulfate-Soluble Proteins. Escherichia coli cells are spun d o w n by centrifugation (30 sec, r o o m t e m p e r a t u r e , 7000 rpm, E p p e n d o r f centrifuge), the cell pellet is r e s u s p e n d e d once with 1 ml of 50 m M Tris-HC1 (pH 7.4), and cells are centrifuged again (30 sec, r o o m
8
VECTORS FOR EXPRESSING CLONED GENES
[1]
temperature, 7000 rpm, Eppendorf centrifuge). The cell pellet is resuspended in 1 × sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer [3 x sample buffer: 62.5 mM Tris-HCl (pH 6.8), 15% (v/v) glycerol, 2.5% (w/v) SDS, 5% (v/v) 2-mercaptoethanol, and 0.001% (v/v) bromphenol blue] and boiled for 5 rain to lyse cells. Hot samples are applied to polyacrylamide gels by using a Hamilton syringe. Proteins Soluble in 8 M Urea. 8 M urea-soluble fraction contains expression products that form so-called inclusion bodies: stable aggregates of partially denatured and partially structured particles, held together mainly by hydrophobic interactions. However, inclusion bodies do not necessarily have to be insoluble. (For a review of solubilization of inclusion bodies and subsequent renaturation see Ref. 15.) Examples for Use of pPLEX Figure 3 describes the expression of a subfragment of rabbit fast skeletal muscle myosin, that is, a 74-kDa portion of light meromyosin (LMM), which is a structural domain of the myosin heavy chain, a component of myosin. The quaternary structure of LMM is assumed to be a coiled coil formed by two molecules.16 The structure of the recombinant LMM74 is similar to that of the native protein, as indicated by electron microscopy. 17 Moreover, recombinant LMM74, like native LMM, can be enriched by high-salt solubilization with 0.5 M KC1 and precipitation by dialysis with low-salt buffers (for details, see the caption to Fig. 3). This property of LMM74 makes it feasible to use this LMM fragment for the generation of fusion proteins with the possible advantages listed above and, in addition, these fusion proteins could be enriched or purified by the high/low-salt method described here [e.g., LMM/human immunodeficiency virus 1 (HIV-1) Tat fusion proteins18]. It should be mentioned that the smaller (-64-kDa) band in Fig. 3 is a product of internal initiation and not a result ofprotease-mediated degradation of LMM74.17 This phenomenon might be of general importance, because it is reasonable to assume that there is no selection pressure against prokaryotic regulatory elements in sequences of higher eukaryotic cells (e.g., cDNA). Other examples for the use of pPLEX to express heterologous open reading frames in E. coli are listed in Table I. 15 R. Rudolph, in "Modern Methods in Protein- and Nucleic Acid Research" (H. Tschesche, ed.), p. 149. de Gruyter, Berlin, 1990. 16 C. Cohen and D. A. D. Parry, Proteins 7, 1 (1990). 17 K. Maeda, G. Sczakiel, W. Hofmann, J.-F. Menetret, and A. Wittinghofer, J. Mol. Biol. 205, 269 (1989). ~8 V. Wolber, K. Maeda, R. Schumann, B. Brandmeier, L. Wiesmiiller, and A. Wittinghofer, Biotechnology 10, 900-904 (1992).
E. coli EXPRESSION PLASMID p P L E X
[1]
A 12
3 4 5 6 7
9
B 1 2
3
4
5
6
7
116 m,, 97 66,,, 4 5 m,-
2 9 m,.
FIG. 3. Expression of a portion of a rabbit fast skeletal muscle light meromyosin (LMM) by use of pPLEX analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting. Bacterial extracts were collected and enriched fractions of recombinant LMM were applied to a 10% (w/v) polyacrylamide gel [U. K. Laemmli, Nature (London) 227, 680 (1970)] and either stained with Coomassie Blue (A) or blotted onto nitrocellulose and reacted first with a polyclonal rabbit anti-myosin antibody, then with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (Sigma, St. Louis, MO) according to the method of Towbin, T. Staehlin, and J. Gordon Proc Natl. Acad. Sci. U.S.A. 76, 4350 (1979) (B). Lanes numbered from 1 to 7 contain the following samples: total lysate of E. coli strain NF1 transformed with pPLEX and grown at 28° (lane 1) and at 42° (lane 2); total lysate of NF1 transformed with pEXLMM 74 (cDNA coding for a 74-kDa portion of the rabbit skeletal muscle LMM inserted into pPLEX) and grown at 28° (lane 3) and after induction for 1, 3, and 5 hr at 42 ° (lanes 4 to 6, respectively); LMM74 after two cycles of high-salt and low-salt buffer as described below (lane 7). Bacterial extracts, that is, soluble proteins, were prepared following the scheme outlined in Fig. 2. For enrichment of expressed LMM (see lane 7) bacteria were harvested by centrifugation after a 5-hr induction at 42 ° and washed once with 50 mM TrisHCI (pH 7.5). Subsequently the cell pellet was lysed. After addition of sodium deoxycholate, KCI was added (final concentration, 0.6 M). The lysate was further incubated for 15 min at room temperature and was centrifuged. The supernatant was dialyzed overnight against 10 mM potassium phosphate (pH 6.5) containing 0.1 M KCI. After dialysis the precipitate was pelleted by centrifugation and dissolved in 10 mM potassium phosphate (pH 6.5) containing 0.6 M KC1. Insoluble proteins were separated again by centrifugation and the cycle was repeated once w;th the supernatant.
l0
VECTORS FOR EXPRESSINGCLONED GENES
[1]
TABLE I pPLEX-DIRECTED EXPRESSIONOF HETEROLOGOUS SEQUENCESIN Escherichia coli Heterologous expression product Wheat Rubisco (ribulose-bisphosphate carboxylase), small subunit Spinach Rubisco activase, two isoforms (41 and 45 kDa) Human papillomavirus type 16, E7 protein Human papillomavirus type 18, E7 protein Dengue virus type 2, nonstructural protein (NS5) Portions of rabbit skeletal muscle light meromyosin (74 and 59 kDa) Portion of human cardiac/3-myosin heavy chain (subfragment l, amino acid residues 1-524)
Detection method, isolation
Reference
Western analysis
a
Purified proteins
b
NF1 and unc 1959 NF1
Western analysis
c
Western analysis
c
N4830-1
Western analysis
d
NFl
Purified protein
e
NFI
Western analysis
f
Host strain N4830-1 and NF1 UT421
M. A. Kaderbhai, M. He, R. B. Beecbey, and N. Kaderbhai, DNA Cell Biol. 9, 11 (1990). b j . B. Shen, E. M. Orozco, and W. L. Ogren, J. Biol. Chem. 266, 8963 (1991). c I. Jochmus and L. Gissmann, personal communication (1991). d A. Bartholomeusz and P. J. Wright, personal communication (1991). e K. Maeda, G. Sczakiel, W. Hofmann, J.-F. Menetret, and A. Wittinghofer, J. Mol. Biol. 205, 269 (1989). f M. Pfordt, Ph.D. thesis, University of Heidelberg, 1991.
C o n c l u d i n g R e m a r k s and Discussion Figure 3 and Table I list e x a m p l e s for the use o f the e x p r e s s i o n plasmid p P L E X . Certainly for p P L E X , and p r e s u m a b l y for o t h e r e x p r e s s i o n vectors as well, there h a v e b e e n a fair n u m b e r o f u n s u c c e s s f u l a t t e m p t s to e x p r e s s h e t e r o l o g o u s p r o t e i n - c o d i n g s e q u e n c e s in E. coli. T h e p a r a m e t e r s generally listed u n d e r T r o u b l e s h o o t i n g (below) are helpful; h o w e v e r , often the s e a r c h f6r i m p r o v e m e n t s r e m a i n s empirical. Critical p a r a m e t e r s for successful p r o d u c t i o n o f p P L E X - e n c o d e d proteins include the c o n d i t i o n s o f induction, that is, the time period and t e m p e r a t u r e o f h e a t shock. B e c a u s e the induction o f the )kpL p r o m o t e r usually is m e d i a t e d b y a t e m p e r a t u r e shift to 42 °, the h e a t - s h o c k r e s p o n s e o f E. coli cells, w h i c h is a c c o m p a n i e d b y i n d u c e d e x p r e s s i o n o f E. coli p r o t e a s e s , 19 c a n affect the stability o f e x p r e s s e d proteins. In addition 19D. W. Mount, Annu. Rev. Genet. 14, 279 (1980).
[1]
E. coli EXPRESSIONPLASMIDpPLEX
11
the time period of induction determines the accumulation of expression products, which has a crucial effect on yields and the physical form of the expression products. In some instances high intracellular concentrations of expressed polypeptides lead to a high potential of formation of insoluble inclusion bodies, whereas low intracellular concentrations result in a higher probability of leaving the expression products in a soluble form. Alternative expression systems for the production of eukaryotic polypeptides (baculovirus and yeast systems, and eukaryotic tissue culture cells) can circumvent some of the fundamental critical points for expression in E. coli as summarized above, particularly posttranslational modifications (e.g., glycosylation). However, expression in E. coli is still one of the most reasonable ways to mass produce structural and enzymatically active polypeptides. One of the more recent improvements of pPLEX was the insertion of additional restriction sites (XbaI, BamHI, SmaI, KpnI, and SstI) between the SalI and BclI sites, creating the modified vector pPLEXI9. 2° Troubleshooting Troubleshooting should include codon usage; secondary structure around start codon (mutations); clonal variability, that is, testing a larger number of transformants; internal translational start sites; and different E. coli strains (e.g., protease-deficient ones, such as unc1857). When heterologous expression products are known to be toxic for E. eoli expression, products can be obtained with expression vector systems that allow almost complete shut-off of the promoter. In this regard the hPc promoter has an advantage over many other widely used promoters, for example, tacp, trcp, or lacp. However, expression systems offering expression cascades (e.g., T7pol-T7 promoter zl) might be alternatives. Acknowledgments We thank B. Miiller for mapping pPLEX restriction sites. 20 j. B. Shen, E. M. Orozco, and W. L. Ogren, J. Biol. Chem. 266, 8963 (1991). 21 F. W. Studier and B. A. Moffatt, J. Mol. Biol. 189, 113 (1986).
12
VECTORS FOR EXPRESSING CLONED GENES
[2]
[2] I n - F r a m e G e n e F u s i o n By HENRIK ~RUM and LARS K. POULSEN Introduction Gene fusion (the joining of unrelated genes) is an extensively used approach in the analysis of a multitude of biological problems. ~To facilitate in vitro gene fusion several vector systems have been developed that carry multiple cloning sites in any of the three reading frames. 2 When the gene of interest has been cloned and sequenced, the desired gene fusion can usually be made by choosing the appropriate vector and restriction site. Alternatively, when there are some means of detecting the gene product, for example, by antibodies, the DNA can be randomly inserted into an expression vector and the clones expressing the desired product identified by subsequent screening with the antibody. Often, neither the sequence of the gene nor an assay for its product is available. In these cases gene fusions can be selected by using vectors known as ORF vectors (open reading frame vectors). 3 ORF vectors utilize the fact that the lacZ gene-encoded fl-galactosidase enzyme is usually active when an additional polypeptide is inserted near its N terminus. Thus, when an open reading frame DNA fragment is inserted near the 5' end of the lacZ gene, the correct fusion (a tripartite gene) will have a Lac ÷ phenotype whereas the incorrect fusions will be L a c - . To confer the Lac ÷ phenotype, the DNA insert must contain an ORF and be in frame with the lacZ gene at both its 5' and 3' ends. Thus, to secure in-frame cloning of a DNA fragment of defined length (i.e., generated by restriction enzyme cleavage) nine different ORF vectors are required. Clearly, handling nine different vectors to make an inframe cloning is impractical. Instead, a single ORF vector is used and in-frame cloning is facilitated by size randomizing the DNA insert prior to cloning. We discuss here a novel in-frame cloning principle that simplifies inframe cloning of DNA fragments of defined length to involve a single vector.
I T. J. Silhavy and J. R. Beckwith, Microbiol. Rev. 49, 398 (1985). 2 p. H. Pouwels, B. E. Enger-Valk, and W. J. Brammer, "Cloning Vectors: A Laboratory M a n u a l . " Elsevier, Amsterdam, 1985. 3 G. M. Weinstock, Genet. Eng. 6, 31 (1984).
METHODS IN ENZYMOLOGY,VOL. 217
Copyright © 1993 by AcademicPress, Inc. All rights of reproduction in any form reserved.
[2]
13
IN-FRAME GENE FUSION
l-"BssXlI-~ rBssHll-- I
r-B,s,X~l -i ]
GCGCGCGCGC CGCGCGCGCG
s ' - cGcGc
GCGCG-S"
REARING F R A M E l
S ' - CGC,;CGC
~CG- S"
REARING FRAME 3
I S' - CGCGCGCGC GCG- S" I I
[,,. REARING FRAME 2
I
i
FIG. 1. The digestion patterns of the BssHlI box. The sequence GCGCGCGCGC [(GC)5] contains three overlapping BssHIl restriction sites, each corresponding to one of the three reading frames. Because cleavage at any one site destroys the two other sites, a particular (GC) 5 box can be cleaved only once.
Materials All chemicals and apparatus referred to in this chapter are commercially available. Except for BssHII [New England Biolabs (Beverly, MA) and Stratagene (La Jolla, CA)] all enzymes were obtained from Boehringer Mannheim (Indianapolis, IN). T4 DNA ligase was purchased in two concentrations [1 unit(U) and 8 U//zl] for use in sticky-end and blunt-end ligation reactions, respectively. Escherichia coli strains DH5a and JM 109 were used as hosts. Principle of Method The restriction enzyme BssHII recognizes and cleaves the alternating sequence GCGCGC and generates 4-b protruding 5' termini. Consequently, the alternating sequence, GCGCGCGCGC [(GC)5], contains three overlapping, mutually exclusive BssHII restriction sites, each corresponding to one of the three different reading frames (see Fig. I). When contained in a vector, the (GC)5 motif is cleaved at an approximate ratio of 2 : 1 : 2, resulting in a mixture of vectors carrying cloning sites in all three reading frames. 4 Thus, by using either one or two (GC)5 boxes, vectors can be 4 H. 0rum and L. K. Poulsen, Nucleic Acids Res. 17, 3107 (1989).
14
VECTORS FOR EXPRESSING CLONED GENES
[2]
/--(-BssHI I) x 3 Mlul / Kpnl / nindllI
/
I I
?
~. % /
T
Pstl
~ I/
plFF 8
SoII/AccI/HIncl i
"Ir
1
hm,,
Smal/Xmal ApaI
~ II~ SP6 1 \
2 8 0 9 bp
A•
\
~
Sacll
\ (B.,.,,). L..NotI
T7promoter EcoRl (BssH ll) x 3 Mlul Kpnl Hindlll Pstl P TRRTACGACTCACTATAGGGCGARTTCAGCGCGCGCGCARCGCGTGGTACCAA GCTrGGCTGCAG I
I
I
I
I
I
I
I
I
!
I
I
!
I
i
I
I
I
I
I
I
I
Sall Xmal Reel Smal Ncol XIncll BamHl Rp,a.l, Secll (DssHll) x 3 Notl GTCGRCGGATCCCCGGGCCCATGGCCGCGGTCGCGTATATGCGCGCGCGCARRGCTGGCGGCCGC I
I
I
!
I
I
i
I
i
I
I
i
!
I
!
I
I
SP6 promoter A GCTI'GAGTATTCTITI'AGTGTG A G CTA A ATAGCTFG 6C GTAATC ATG GTC AT I
I
I
I
I
|
I
I
I
|
I
I
I
I
I
I
|
I
!
I
!
I
'~ Plac
I
lacZ Initiation I ¢odon I
FIG. 2. Schematic representation of the ORF vector plFF8. The sequence of the 5' part of the lacZ-c~ gene including the lacZ initiation codon, the ORF multiple cloning site, the two (GC)s boxes, the NotI site, and the SPG and T7 promoters are shown. Pt,c designates the lacZ promoter. The vector carries the /~-lactamase gene (bla), conferring resistance to ampicillin.
constructed that allow in-frame cloning of DNA fragments of defined length at one or two fusion points, respectively. Figure 2 shows a schematic representation of an ORF vector, termed pIFF8 (in-frame fusion) constructed by this principle. It is derived from a previously described vector, pIFF5, 4 and carries the inducible lacZ gene
[2]
IN-FRAME GENE FUSION
15
promoter and the lacZ a fragment. Inserted near the 5' end of the lacZ ct fragment are two (GC) 5 BssHII recognition boxes that allow cleavage randomization at two cloning points. In the previous vector, plFF5, a 1.2kilobase (kb) fragment was inserted between the two (GC)5 boxes to avoid the possibility that close proximity of these boxes would prevent cleavage at some of the BssHII sites. In plFF8 this spacer fragment is replaced by a multiple cloning site that has two features: (1) it does not contain stop codons in any of the three reading frames; and (2) it restores the lacZ reading frame, giving the plFF8 vector a Lac ÷ phenotype. As discussed in Procedure 1.2 (below), these new features facilitate in-frame cloning by an indirect procedure. To allow verification of selected clones, the plFF8 vector further carries a unique NotI site located upstream of the 5'-most (GC)5 box (Section 4). Furthermore, the vector contains an SP6 and T7 promoter sequence that allows transcription through the multiple cloning site. Methods 1. Preparation of Vector for In-Frame Cloning The pIFF8 vector can be used to select open reading frames in DNA/ cDNA fragments carrying BssHII-compatible sticky ends (fragments generated by BssHII and/or MluI cleavage) or blunt ends. To prepare the vector for either type of cloning, it is first cleaved with BssHII to produce the nine possible cloning combinations. It is important that the vector sample is totally cleaved at this step because residual uncleaved pIFF8 will give rise to false positives in subsequent transformation/plating (both the desired recombinant pIFF8 vector as well as the pIFF8 vector itself have a Lac + phenotype). After cleavage with BssHII, the vector is treated with calf intestinal phosphatase (CIP) to remove the terminal 5'-phosphate groups. This treatment prevents the vector from recircularizing without insert and thus eliminates yet another source of false positives ( - 3 5 % of recircularized pIFF8 vectors will have a Lac + phenotype). If blunt-ended DNA fragments are to be cloned, the vector is further treated with Klenow polymerase in the presence of all four dNTPs to fill in the BssHII sticky ends. Although very little vector is used in a cloning experiment, it is convenient to prepare an excess amount that can subsequently be stored as a "ready to use" vector. The following procedures will usually give a good result. Procedure 1.1: BssHII Cleavage. Mix 10 tzg of vector (purified by CsCI gradient centrifugation) and 50 U of BssHII enzyme in a 100-tzl reaction
16
VECTORS FOR EXPRESSING CLONED GENES
12]
containing 25 mM NaCI, 6 mM Tris-HCl, pH 7.4, 6 mM MgCI2, and 5 mM dithiothreitol (DTT). Overlay the reaction with a drop of paraffin oil and incubate 3 hr at 50 °. Place the reaction on ice; remove a 3-/zl aliquot (-0.3 /zg of vector) and analyze the extent of cleavage by electrophoresis through a 1% TAE (tris-acetate-ethylenediaminetetraacetic acid) agarose gel using appropriate DNA size markers. If more than one vector band is observed, add more BssHII enzyme and continue the incubation. When the cleavage is complete, extract twice with phenol and chloroform and precipitate the DNA with 1/10 vol of 2.5 M sodium acetate, pH 5.2, and two vol of 96% ethanol for 30 min at - 2 0 °. Recover the DNA by centrifugation at 12,000 g for 30 min at 4 ° and redissolve in 20 ~1 of TE [I0 mM Tris CI, pH 8.0 and I mM ethylenediaminetetraacetic acid (EDTA)]. Procedure 1.2: Phosphatase Treatment. Mix the BssHII-cleaved vector and 3 U of CIP enzyme in a 50-/xl reaction containing 50 mM Tris-HCl, pH 9.0, 1 mM MgCI 2, 1 mM ZnC12, and 1 mM spermidine. Incubate at 37 ° for 30 min, add an additional 2 U of CIP enzyme, and continue the incubation for another 30 min. Add 5 tzl of STE buffer (100 mM Tris C1, pH 8.0, 1 M NaCI, and I0 mM EDTA), 5/zl of 10% (w/v) sodium dodecyl sulfate (SDS), and 40 ~1 of distilled water. Incubate 15 min at 70 °, extract with phenol and chloroform, and precipitate the DNA as above. Redissolve the vector in TE buffer to a final concentration of 50 ng//zl (usually between 150 and 200/~1). Store the vector at - 2 0 ° to prevent evaporation. Note: It is advisable to test the efficiency of the dephosphorylation step by trying to recircularize the vector in the absence of added insert. To do this, set up a "vector alone" standard ligation (Section 2, procedure 2.1 without target DNA) and transform and plate competent E. coli cells as described in Section 3, procedure 3.1. Optimally, there should be no colonies on the plate. Usually, however, even properly dephosphorylated vector preparations give rise to several colonies. If there are many colonies on the plates and these are predominantly blue (Lac+), the problem relates to incomplete cleavage with the BssHII enzyme (uncleaved plFF8 vector is Lac÷). In contrast, if the main part of the colonies are white (Lac-), the problem relates to the dephosphorylation step (about 65% recircularized vectors are Lac-). Procedure 1.3: Filling-In Reaction. Mix I0/zg of BssHII-cleaved/dephosphorylated vector and 5 U of Klenow polymerase in a 100-/zl reaction containing 50 mM Tris-HC1, pH 8.0, 10 mM MgCI 2 , 100 mM NaCI, and 0.5 mM dATP, dCTP, dTTP, and dGTP. Incubate at 23-25 ° for 30 min, heat to 70 ° for 10 min, extract with phenol and chloroform, and precipitate the vector as above. Dry and redissolve the vector in TE buffer to a
[2]
IN-FRAME GENE FUSION
17
final concentration of 100 ng//zl (-75-100/xl). Store at - 2 0 ° to prevent evaporation.
2. Preparation and Cloning of DNA Fragments Any of several reliable methods can be used to prepare DNA/cDNA for cloning in pIFFS. For direct in-frame cloning in pIFF8 (prepared as above) the foreign DNA fragment must carry either BssHII-compatible ends or blunt ends. There are presently only two enzymes (BssHII and MluI) that will provide BssHII-compatible sticky ends and the recognition sequence for the BssHII enzyme appears to be rare in DNA. Consequently, it may not be possible to locate the gene of interest to a BssHII and/or MluI-generated DNA fragment of a size suitable for cloning. In contrast, there is a whole range of enzymes that will generate blunt ends and as such it will usually be possible to locate the gene of interest in a properly sized blunt-ended fragment. Alternatively, the DNA fragment of interest can be cloned in frame by a simple, indirect procedure. First, the DNA fragment is cloned into one of the several unique restriction sites in the ORF multiple cloning site of pIFFS. This cloning destroys the lacZ reading frame, allowing the desired recombinant to be selected by its L a c - phenotype. Second, purified vector from the selected L a c - clone is cleaved with BssHII, extracted with phenol and chloroform, precipitated with ethanol, and religated using T4 DNA ligase (Section 2, procedure 2.3). This ligation shuffles the BssHIIexcised insert/vector fragments, with the result that a subset of inserts are brought in frame with the lacZ gene. Clones containing these vectors can then be selected by their Lac + phenotype. When using the indirect inframe cloning procedure as outlined above, self-circularization of vectors during the shuffling step will produce a background of false positives, that is, Lac + vectors without insert. As described in the following section, these vectors can often be distinguished from the desired recombinants by the intensity of the blue color of the resulting colonies. Alternatively, the BssHII-excised insert can be isolated by agarose gel electrophoresis and cloned in a premade pIFF8 vector (Section l) to avoid the selfcircularized vector background. The optimal conditions for ligating vector/DNA fragments carrying blunt ends or sticky ends are somewhat different. We usually obtain a good result using the following conditions.
Direct In-Frame Cloning Procedure 2.1.: Ligation of Vector and DNA Fragments with Sticky Ends. Mix 100 ng of prepared vector (Section l) and target DNA in a molar
18
VECTORS FOR EXPRESSING CLONED GENES
[2]
ratio of 1 : 3 with I U of T4 DNA ligase (1 U//A) in a 20-~1 reaction containing 50 mM Tris-HC1, pH 7.8, 5 mM MgCI2, 1 mM ATP, 20 mM DTT, and 50/.tg/ml of bovine serum albumin (BSA). Incubate for 4-16 hr at 16° and store at - 2 0 ° until use. Procedure 2.2: Ligation of Vector and DNA Fragments with Blunt Ends. Mix 200 ng of prepared vector (Section 1) and target DNA in a molar ratio of 1 : 3 with 12 U of T4 DNA ligase (8 U//A) (efficient blunt-end ligation requires a great deal ofT4 DNA ligase) in a 20-/xl reaction containing 50 mM Tris-HCl, pH 7.8, 5 mM MgCI 2 , 1 mM ATP, 20 mM DTT, 5% (w/v) polyethylene glycol (PEG) 6000, and 50/xg/ml BSA. Incubate for 4-16 hr at 23-25 ° and store at - 2 0 ° until use.
Indirect In-Frame Cloning Procedure 2.3: Shuffling of Vector and Insert. Mix 100 ng of BssHIIcleaved vector obtained from a L a c - clone with 1 U of T4 DNA ligase (1 U//xl) in a 20-/zl reaction containing 50 mM Tris-HCl, pH 7.8, 5 mM MgCI2, 1 mM ATP, 20 mM DTT, and 50/zg/ml of BSA. Incubate for 4-16 hr at 16° and store at - 2 0 ° until use. 3. Transformation and Selection of Recombinants The plFF8 vector uses lacZ a complementation to produce a Lac ÷ phenotype and therefore requires E. coli strains carrying the lacZ AM15 gene as host; that is, E. coli DH5a, XLl-blue, JM101-109, etc. To prepare the cells for transformation, we use the CaC1 method 5 or, where improved transformation efficiencies are required, the method of Hanahan. 6 Procedure 3.1: Transformation. Mix 100/xl of competent cells and 2/A of the ligation reaction in an Eppendorf tube and incubate on wet ice for 1 hr. Place the tube in a water bath at 42 ° for 45 sec and return the tube to the wet ice for 2 min. Add 900/~1 of LB medium [1% (w/v) Bactotryptone, 0.5% (w/v) Bacto-yeast extract (Difco, Detroit, MI) 1% (w/v) NaCI, pH 7.5] and incubate the tube at 37° for 1 hr in a shaking incubator. To facilitate subsequent isolation of individual clones, plate 20- and 200/A samples onto LB agar plates containing 0.5 mM isopropyl-~/-o-thiogalactopyranoside (IPTG), 40/zg/ml 5-bromo-4-chloro-3-indolyl-fl-D-galactopyranoside (X-Gal) and 50/zg/ml of ampicillin. Incubate the plates (head up) overnight at 37 °. Following transformation and plating, one should in principle obtain 5 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2nd Ed., pp. 182-184. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989. 6 D. Hanahan, in " D N A Cloning" (D. M. Glover, ed.), Vol. 1, p. 109. IRL Press, Washington, D.C., 1985.
[2]
IN-FRAME GENE FUSION
19
(1) white colonies (Lac-) containing vectors with either a DNA fragment that is not an ORF or an ORF DNA fragment inserted out of frame with the lacZ gene and (2) blue colonies (Lac ÷) containing vectors with a correctly fused ORF DNA. Unfortunately, it is rarely as simple as that. Thus, blue colonies may also contain (1) a vector with a DNA fragment that is not in frame with the 5' end of the lacZ gene but contains a translation initiation site in frame with the 3' part of the lacZ ~ gene or (2) vectors without a DNA insert (caused by either insufficient BssHII cleavage or vector self-circularization). To show that an inserted DNA fragment contains an ORF, it must therefore subsequently be verified that Lac" vectors contain the DNA insert and that translation initiates at the lacZ translation start site. In selecting a number of candidate clones for these analyses the color of the colonies may be of some help. Thus, depending on how the insert DNA affects transcription and/or translation of the tribrid gene/mRNA, affects folding and stability of the tribrid protein, and so on, colonies containing the correct fusion will range in color from deep to light blue (for a detailed discussion of factors affecting clonal color development, see Ref. 3). Similarly, the blue color of colonies containing a vector where translation initiates within the insert will depend on several factors, including how efficiently translation initiates within the insert. In contrast, Lac ÷ colonies containing vectors without a DNA insert will always be deep blue. Thus, when colonies exhibit different shades of blue one usually selects a number of clones from each group for further verification, with preference for those that are light blue. 4. Verification of Selected Clones To verify that the inserted DNA contains an ORF, the first step is to prepare a vector minipreparation from each of the selected clones. For this we use the alkaline lysis method, which is both rapid and reliable. 7 Next, purified vectors are digested with restriction enzymes followed by electrophoresis in agarose or polyacrylamide gels to determine which of the vectors contain the correct insert. Digestion with BssHII excises the insert, thus allowing its size to be determined against coelectrophoresed DNA size markers. However, BssHII is an expensive enzyme and for this reason the use of alternative restriction enzymes should be considered. For instance, combined digestion with EcoRI and NotI also excises the insert. Alternatively, when the size of the insert is such that vector plus insert can be distinguished from vector alone, any restriction enzyme that cleaves only once in the vector can be used (in this case several fragments 7 H. C. Birnboim and J. Doly, Nucleic Acids Res. 7, 1513 (1979).
20
VECTORS FOR EXPRESSING CLONED GENES
[2]
A Notl
P~.c ~
i:~c'::~c~] INSERT llacZ-a ~-RTG
I
(RTG)
'~................. ...................................
B
TRIBRID GENE
•
mRNA
~-Lac +
TRRNSLATI ON STARTS WITHIN INSERT
... Lac +
TRANSLATION STARTS RTlacZ 5" END
J DIGESTWITHNotiFILL IN TXE STICKY E N D S - RELIGATE.
Insertion 4bp.
of
Plac ~ C ~ c S ~ C C c S C . ~
INSERT
~" RTG
i lacZ
(ATG)
- ~
I
TRIBRIO GENE
i~
mRNA
i !
: .............
.... ~
-
~
"
"~
Lac'l"
Lac-
TRANSLATION STARTS WITHIN INSERT TRANSLATION STARTS ATIacZ 5" ENn
FI6.3. Schematic outline of the strategy to distinguish between translation initiation at the lacZ translation start site or from within the insert. (A) The recombinant vector isolated from a Lac + clone; translation initiation at the lacZ translation start site or from within the insert both confers a Lac + phenotype on the host. (B) Introduction of 4 bp between the lacZ translation initiation site and the 5' end of the insert disrupts the lacZ reading frame read from the lacZ translation initiation site, but does not affect the reading frame initiated from translation start sites within the insert. Thus the phenotype of vectors containing a correctly fused insert where translation initiates at the lacZ start site will change from Lac + to Lacwhereas incorrectly fused inserts will remain Lac +.
may result from the digestion depending on whether target sequences for the enzymes are present in the insert or not). Those vectors that contain the correct DNA insert are then analyzed to distinguish between the possibility that translation initiates at the lacZ translation start site or within the insert. The rationale behind this analysis is shown schematically in Fig. 3. First, the selected vectors are digested with NotI. Provided there are no NotI sites in the insert (NotI recognizes an 8-bp DNA sequence and its target sequence is thus rare in DNA), this digestion linearizes the vector between the lacZ translation start site and
[2]
IN-FRAME GENE FUSION
21
the 5' end of the DNA insert. The NotI site is then filled in with Klenow polymerase in the presence of all four dNTPs (Section 1, procedure 1.3) and the vector is recircularized using T4 DNA ligase (the ligation reaction is similar to procedure 2.3 except that the reaction volume is increased to 100/xl to favor vector self-circularization). This treatment introduces 4 bp between the 5' end of the lacZ gene and the insert, thereby disrupting the lacZ reading frame read from the lacZ translation start site. In contrast, the lacZ reading frame read from any spurious translation start site within the insert is not affected. Thus, the phenotype of vectors containing the correctly fused insert will change from Lac + to Lac , whereas vectors that do not will remain Lac +. Examples
Selection of Open Reading Frames in DNA/cDNA The pIFF series of vectors, and in particular pIFF8, are recent vector constructions and examples on experimental applications are therefore limited at present. The pIFF5 vector has been used to select the ORF in a 1.6-kb cDNA fragment encoding an internal part of the enzyme phenylalanine ammonia-lyase from the basidiomycete yeast Rhodotorula glutinis.4 In a parallel experiment the corresponding genomic pal gene fragment did not contain an ORF, as evidenced by the lack of blue colonies, and this was later shown to be due to the presence of several small introns. 8 The major difference between pIFF8 and pIFF5 is in the spacing of the two (GC)5 boxes. In pIFF5, these boxes are separated by a 1.2-kb spacer fragment whereas in pIFF8 they are separated by a small ORF multiple cloning site. To determine whether the decrease in spacing between the two (GC)5 boxes in pIFF8 would affect the pattern of BssHII cleavage, pIFF8 was cleaved with BssHH and the small multiple cloning site fragment purified from a 2% (w/v) agarose gel. The purified fragment was then dephosphorylated with calf intestinal phosphatase, labeled with [7-32p]ATP and T4 DNA kinase, and digested with PstI, which cleaves the labeled fragment into two unequal halves. Finally, the labeled products were separated by electrophoresis in a polyacrylamide sequencing gel and autoradiographed. Two sets of bands corresponding to cleavage at all BssHII sites in the 3'-most (GC)5 box [31, 33, and 35 nucleotides (nt)] and 5'-most (GC)5 box (50, 52, and 54 nt) were detected on the film, showing that all BssHII sites in both (GC) 5 boxes were accessible to cleavage. Moreover, the bands corresponding to 33 and 52 nt were less intense than 8 j. G. Anson, H. J. Gilbert, J. O. Oram, and N. P. Minton, Gene 58, 189 (1987).
22
VECTORS FOR EXPRESSING CLONED GENES
[2]
the bands corresponding to 31, 35, 50, and 54 nt, supporting the previous observation that the center BssHII site in a (GC)5 box is cleaved less frequently than the flanking BssHII sites. From this we conclude that the close proximity of the two (GC)~ boxes in plFF8 does not have any major distortive effect on BssHII cleavage characteristics compared to the characteristics of previously described plFF vectors. Discussion Applications and Limitations
This chapter has focused on the use of (GC)5 boxes in the construction of gene-fusion vectors. When applied to ORF vectors the system offers the major advantage that DNA fragments generated by restriction enzyme cleavage can be cloned in frame without the need for prior size randomization. In addition to simplifying the use of ORF vectors in general, this feature potentially expands their uses. For instance, information on possible introns and a rough map of protein-coding domains in a cloned gene can be rapidly provided by subcloning specific restriction fragments in plFF8 and such information may be useful in setting up a sequencing strategy. Likewise, provided a correctly fused ORF DNA insert is sufficiently large so that the ORF can be considered biologically significant, the reading frame can be established by sequencing through the vector/ insert junctions, and this knowledge is useful in subsequent interpretation of sequencing data. As with other ORF vectors, the proper function of plFF8 requires that the lacZ-encoded part of the fusion protein retain enzymatic activity. In plFF8, the foreign DNA is inserted into the small lacZ ~ gene, which must successfully complement the host encoded product of the lacZ AM 15 gene to produce the Lac ÷ phenotype. Thus, compared to other ORF vectors that usually carry the entire lacZ gene, it may be expected that insertion of foreign ORF DNA fragments in plFF8 has a more pronounced effect on the Lac ÷ phenotype. Consistent with this notion, insertion of a 1.6-kb pal cDNA fragment in plFF8 produced light blue colonies that, however, turned deep blue when the lacZ t~ gene was substituted by the entire lacZ gene. This suggests that the functional limits of plFF8 can be expanded by insertion of the entire lacZ gene. On the other hand, when using the indirect in-frame cloning procedure, the present construction is probably advantageous in that a clear effect of the DNA insertions on the Lac ÷ phenotype allows an easier distinction between the desired recombinants and the false positives (i.e., vectors without insert).
[3]
HETEROLOGOUS
PROTEIN PRODUCTION
BY
B. brevis
23
[3] H i g h - L e v e l S e c r e t i o n o f H e t e r o l o g o u s P r o t e i n s b y Bacillus brevis By SHIGEZO U D A K A a n d H I D E O YAMAGATA Introduction
Among various host-vector systems for the production of foreign proteins in microorganisms, the use of Bacillus breois as a host offers the advantage that proteins are secreted directly into the culture medium, where they are accumulated at high levels in a relatively pure state. The secreted proteins are usually correctly folded, soluble, and biologically active. Bacillus brevis is known to be a harmless inhabitant of soil, milk, and cheese. Many of these advantages are shared with another thoroughly studied Bacillus species, B. subtilis. The major advantage of B. brevis over B. subtilis, however, is a very low level of extracellular protease activity, so that secreted proteins are usually stable and not significantly degraded. 1 For example, human a-amylase was secreted in quantities of up to 60 mg/liter by B. brevis, 2 whereas none was produced by B. subtilis. 3 Bacillus brevis 47 was isolated from soil as a protein-hyperproducing bacterium and was found to show little extracellular protease activity. 1.4 The two main proteins secreted by B. brevis 47 were indistinguishable from the two major proteins found in the outer two protein layers of the cell wall. The major cell wall proteins (CWP) synthesized during the logarithmic phase of growth form hexagonal arrays on the cell surface. During the early stationary phase of growth, the protein layers begin shedding concomitantly with a prominent increase in protein secretion. 5 During the stationary growth phase, cells continue to synthesize and secrete the cell wall proteins. These proteins do not stay on the cell surface, but instead accumulate in the medium as extracellular proteins with concentrations up to 20 g/liter of culture. The amount of extracellular protein reaches more than twice that of intracellular proteins. The genes coding for the major cell wall proteins (an outer wall protein and a middle wall I H, Takagi, K. Kadowaki, and S. Udaka, Agric. Biol. Chem. 53, 691 (1989). 2 H. Konishi, T. Sato, H. Yamagata, and S. Udaka, Appl. Microbiol. Biotechnol. 34, 297 (1990). 3 T. Himeno, T. Imanaka, and S. Aiba, FEMS Microbiol. Lett. 35, 17 (1986). 4 S. Udaka, Agric. Biol. Chem. 40, 523 (1976). 5 H. Yamada, N. Tsukagoshi, and S. Udaka, J. Bacteriol. 148, 322 (1981).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
24
VECTORS FOR EXPRESSING CLONED GENES
[3l
protein) were cloned, and an operon (cwp) for cell wall protein genes was f o u n d . 6,7
Taking advantage of these characteristics of B. brevis, we developed a host-vector system for efficient production of heterologous proteins. The 5' region of the cell wall protein gene containing the powerful promoter and the signal peptide-coding sequence is utilized to construct expression-secretion vectors that are introduced into the protein-hyperproducing B. brevis. Media and Reagents
T2U medium contains 10 g of polypeptone (Nihon Pharmaceutical, Tokyo, Japan; tryptone, Difco, Detroit, MI), 5 g of meat extract (Wako Pure Chemical Industries, Osaka, Japan), 2 g of yeast extract (Difco), 0.1 g of uracil, and 10 g of glucose per liter. PM medium contains 20 g of polypeptone, 10 g of meat extract, 4 g of yeast extract, 0.1 g of uracil, and 10 g of glucose per liter and 2 mM CaC12 (pH is adjusted to 7 with NaOH). Solid medium contains 15 g of agar per liter. Erythromycin (10/zg/ml) or neomycin (60/xg/ml) is added for the growth of plasmid-bearing bacteria. MTP is prepared as follows: 20 ml of 0.1 M sodium maleate (pH 6.5), 10 ml of phosphate buffer [7% (w/v) K2HPO4 and 2.5% (w/v) KH2PO4)], and 18 ml of H20 are mixed and sterilized by autoclaving; after cooling the mixture, 2 ml of 1 M MgCI2 and 50 ml of T2U medium are added. Polyethylene glycol (PEG) solution is prepared by dissolving 40 g of PEG 6000 (average M r 7500) in 20 ml of 0.1 M sodium maleate (pH 6.5) and adjusting the volume to 100 ml with H20. TE contains 10 mM Tris-HCl (pH 8) and 1 mM disodium salt of ethylenediaminetetraacetic acid (EDTA). Sterilization of all the solutions, except for antibiotics, is carried out by autoclaving at 120° for 15 min. Host Bacterium Bacillus brevis 47-5Q is derived from strain 47-5, which is a uracilrequiring mutant of the wild-type 47. 4 Strain 47-5Q generally shows one or two orders of magnitude higher transformability and certain plasmids are more stably maintained in this strain than in strain 47-5. Bacillus brevis 47-5Q shows little protease activity in its culture supernatant. This bacterium hardly sporulates when cultured in ordinary media. 6 H. Yamagata, T. Adachi, A. Tsuboi, M. Takao, T. Sasaki, N. Tsukagoshi, and S. Udaka, J. Bacteriol. 169, 1239 (1987). 7 S. Tsuboi, R. Uchihi, T. Adachi, T. Sasaki, S. Hayakawa, H. Yamagata, N. Tsukagoshi, and S. Udaka, J. Bacteriol. 170, 935 (1988).
[3]
HETEROLOGOUS PROTEIN PRODUCTION BY B. brevis
25
Preservation of Bacteria Bacillus brevis cells, including those having plasmids, can be preserved at or below - 80° in the presence of 20% (w/w) glycerol for several years. Routinely, they may be maintained at room temperature on plates (T2 agar is appropriate) by transferring every 2-3 weeks. The bacteria will die at 4°. It is advisable to keep cells harboring plasmids containing foreign genes at or below - 80°, becuase both plasmids and hosts tend to mutate so that they no longer produce the foreign proteins.
Plasmids pUB 1108 is a high-copy-number plasmid in B. brevis, useful for overproduction of polypeptides from cloned genes. The neomycin resistance gene on this plasmid can be used as a selective marker for transformation. However, B. breois 47 spontaneously gives rise to mutants resistant to this drug at a relatively high frequency, so that examination for the presence of the plasmid is necessary to distinguish transformants from the spontaneous mutants. pHWl 9 is a low-copy-nu.mber plasmid and is useful as a cloning vector, especially when products of the cloned gene are deleterious to the host cells. Therefore, pHW 1 was used for cloning the genes encoding the middle wall protein (MWP) and the outer wall protein (OWP) ofB. brevis 47. The erythromycin resistance gene (Em0 on this plasmid, originally found in pE 194, is useful for selection of transformants because almost no spontaneous erythromycin-resistant mutants appear under the standard transformation conditions, pRU100 was constructed by inserting a multicloning site derived from M13mpl9 between the EcoRI and PvulI sites of pHWl. Another series of vectors was constructed from a low-copy-number cryptic plasmid, pWT481, found in B. brevis 481. l° pHY481 was constructed by inserting the erythromycin resistance gene into pWT481 and is stably maintained in B. brevis 47 even in the absence of the selective drug. H Although pHY481 and its derivatives have not been used extensively to date, results suggest that these plasmids are useful for efficient protein production. 6 8 T. McKenzie, T. Hoshino, T. Tanaka, and N. Sueoka, Plasmid 15, 93 (1986). 9 S. Horinouchi and B. Weisblum, J. Bacteriol. 150, 804 (1982). 10 H. Yamagata, W. Takahashi, K. Yamaguchi, N. Tsukagoshi, and S. Udaka, Agric. Biol. Chem. 48, 1069 (1984). fl H. Yamagata, K. Nakagawa, N. Tsukagoshi, and S. Udaka, Appl. Environ. Microbiol. 49, 1076 (1985).
VECTORS FOR EXPRESSING CLONED GENES
26
[3]
60
Mfl I
ATCAGATCCGCTATCCTOTCTTACAACTTOOCTOTTOTAAACTTTOAAAATOCATTAOOA 120
AATTAACCTAATTCAAGCAAGATTATOAO(]TTTT(]AACCAAATTGGAAAAAGOTTCAGTC
l~ 18o 0TGACAGGCCGCCATATOI'CCCCTATAATACGGATTOTGGCOGATGTCACTTCOTACATA 240 ATGGACAGOTGAATAACGAACCACGAAAAAAACTTTAAATTTTTTTCGAAGGGGCCGCAA
Z~ 300 CTTTTOATTCGCTCAGGCOTTTAATAGGATOTGACACGAAAAACOGGOAAT~rOTOTAAAA EcoRI SpeI 3~ 360 AAOATTCACGAATTCTAGCAC, TTGTGTTACACTAGTGATTGTTGCATTTTACACAATACT
41 5~ SDI 4zo GAATATACTAOAGATTTTTAACACAAAAAGCGAGGCTTTCCTGCGAAAGGAGGTGACACG 480 COCTTGCAGGATTCGGGCTTTAAAAAGAAAGATAGATTAACAACAAATATTCCCCAAGAA fHetGlnAspSerGlyPhebysLysLysAspArgLeuThrThrAsnlleProGlnGlu S D 2 Hpal 540 CAATTTGTTTATACTAGAGGAGGAGAACACAAGGTTATGAAAAAGGTCGTTAACAGTGTA GlnPheValTyrThrArgG-G-f~yGi'uIllsgysVa i Me t bys Lys ValValAsnSer Val
ApaLI Ncol PstI B a m H I Sall TTGOCTAOTOCACTCGCACTTACTGTTOGTCCCATOUGTTTCOCTGCAGGATCGOTCGAC beuA l a S e r A l a b e u A l m b e u T h r V a l A l a P r o M e t A l a P h e A l a
,l.XcTb:IGA~PTA~CABG~ITICITCXTIC°~Gg: i~'~77~CIT~
"
EcoRI,,
/
/<
FIG. 1. Structure of the expression-secretion vector pNU210. The closed bar indicates the 5' region of the mwp gene containing multiple promoters and the signal peptide-coding sequence. The open bar indicates a multicloning site (MCS). The DNA and amino acid
[3]
HETEROLOGOUS PROTEIN PRODUCTION BY B. brevis
27
Preparation of Plasmid DNA from Bacillus brevis Cells The method of Birnboim and Doly 12 can be used to obtain plasmid DNA. This method can be scaled up to obtain large amounts of plasmid DNA and DNA can be purified by CsCl-ethidium bromide centrifugation 13 or adsorption to glass powder in a high-salt solution and elution with water (e.g., The GeneClean II kit; Bio 101, San Diego, CA). Construction of Expression-Secretion Vector, pNU210 As described above, the cell wall proteins (OWP and MWP) are synthesized and secreted into the medium efficiently, even during the stationary phase of growth in B. brevis 47. This suggested that the 5' region of the cwp operon would greatly facilitate the expression of downstream heterologous genes and secretion of the gene products. A 600-bp AluI fragment containing the five tandem promoters, dual translation initiation sites, and the MWP signal peptide-coding region6,14was isolated and used to construct expression-secretion vectors. The structure of the one such expression-secretion vector thus far constructed, pNU210, is shown in Fig. 1. pNU210 is a multicopy plasmid with the replication origin of pUB 110 and the Em r gene of pHWI. The multicloning site on the plasmid is convenient for the insertion of foreign genes to construct transcriptional fusion with the cwp operon or translational fusion with the 5' terminal portion of the mwp gene. In the latter case, the gene product will be fused to the MWP signal peptide. An ApaLI or NcoI site located within the MWP signal peptide-coding region is useful for constructing transcriptional or translational fusions of the MWP gene with foreign genes (see Fig. 1). By inserting the appropriate synthetic DNA fragment encoding the COOH-terminal portion of the MWP signal peptide between the ApaLI or NcoI site and the foreign gene, the foreign proteins directly fused with the MWP signal peptide can be 12 H. C. Birnboim and J. Doly, Nucleic Acids Res. 7, 1513 (1979). ~3 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989. 14T. Adachi, H. Yamagata, N. Tsukagoshi, and S. Udaka, J. Bacteriol. 171, 1010 (1989).
sequences of these regions are shown in upper part of the figure. Vertical arrows along the top of the DNA sequence indicate transcription start sites, 1 to 5. SD1 and SD2 are the ribosome-binding sites located upstream of the dual translation initiation sites (TTG at nucleotides 424 to 426 and ATG at nucleotides 517 to 519). The signal peptide-coding sequence is underlined. All restriction sites shown here, except HpaI, NcoI, MflI, and SacI, are unique to pNU210.
28
VECTORS FOR EXPRESSING CLONED GENES
(Nrul) pBR322
ApaLI
NcoI
[3]
PstI BamHI SalI
AGTGCACTCGCACTTACTGTTGCTCCCATGGCTTTCGCTGCAGGATCCGTCGAC TCACGTGAGCGTGAATGACAACGAGGGTACCGAAAGCGACGTCCTAGGCAGCTG
AlaPheAla
MWP Signal P e p t i d e ~ ~
,MCS |
Cleavage site
XbaI
KpnI
BglII XhoI
SacI
HindIII
EcoRI
TCTAGAGGTACCAGATCTCTCGAGGAGCTCAAGCTT pBR3 2 2 AGATCTCCATGGTCTAGAGAGCTCCTCGAGTTCGAA
MCS
/
FIG. 2. Structure of pBR-AN3, used for making subconstructs in E. coli. The DNA having the nucleotide sequence shown is inserted between the NruI and HindlII sites of pBR322. The sequence from the 5' terminus to the Pst! site encodes the COOH-terminal portion of the MWP signal peptide. Downstream from the cleavage site of the signal peptide is a multicloning site (MCS), the same as that inserted in pNU210 (Fig. 1). The ampicillin resistance gene on pBR322 was replaced by that of pUC18. Therefore, all restriction sites shown above the DNA sequence except ApaLI are unique to pBR-AN3.
synthesized and processed efficiently. This results in accumulation in the medium of the foreign proteins with no additional amino acid residues at their NH2 termini. A PstI site located at the cleavage site of the MWP signal sequence can also be used for the production of foreign proteins with the correct N H 2 terminus. To express genes efficiently, transcription termination is presumed to be important. However, when certain terminator DNA fragments, such as that of a bacterial a-amylase gene, were inserted downstream from the foreign gene, no marked difference in gene expression was observed. Construction of Plasmid pBR-AN3 for Insertion of Foreign Genes To construct a plasmid for the production of the heterologous protein in B. brevis, an Escherichia coli vector, pBR-AN3, useful for linking the foreign gene with the cell wall protein (MWP) signal sequence ofB. brevis 47, was prepared (Fig. 2). Construction procedures are as follows. 1. Insert the foreign gene between the NcoI or PstI site and one of the multicloning sites (e.g. HindIII) in pBR-AN3. The direction of transcription for the inserted DNA must be from the NcoI site to the HindIII site. In the case of the NcoI site, a synthetic linker may be used to connect the
[3]
HETEROLOGOUS PROTEINPRODUCTIONBY B. brevis
29
MWP signal sequence and the DNA encoding foreign mature protein. This connection must be done so that the fused gene encodes exactly the same amino acid sequence as the mature foreign protein directly following the MWP signal peptide. 2. Introduce the plasmid DNA thus prepared into competent cells of E. coli HB 101 (or any other appropriate strain) according to the standard method. 13 Transformants are selected on LB agar plates containing 50 ~g/ml ampicillin following incubation at 37° for one day. 3. Purify the plasmid DNA from the transformant and digest it with A p a L I and any enzyme of the multicloning site on pBR-AN3. After electrophoresis on agarose gel, elute the DNA fragment containing the m w p foreign gene fusion and purify it by means of a glass powder method (e.g., The Gene Clean II kit; Bio 101, San Diego, CA) or electrophoresis. 4. Ligate the DNA thus obtained to the large fragment of pNU210 generated by digestion with A p a L I and the enzyme used to cut the 3' flanking region of the foreign gene. Use the ligated DNA to transform B. brevis 47-5Q. From among transformants, select clones that produce foreign proteins.
Transformation of Bacillus brevis
Tris-Polyethylene Glycol Method The original method 15 was modified as follows. 1. Take onto a toothpick a small amount ofB. brevis 47-5Q cells grown on a T2U plate, and inoculate it into 5 ml of T2U medium and grow overnight with shaking at 37° . 2. Dilute the overnight culture 100-fold in 5 ml of the same medium and incubate at 37° with vigorous shaking for 4 to 5 hr. Alternatively, suspend fresh cells grown overnight on a T2U plate in 5 ml T2U medium with an initial OD660of approximately 0.05 and incubate as above. 3. The following steps must be done at room temperature. At the logarithmic phase of growth (from middle to late, i.e., when the OD660 is 1.0-1.7), collect cells in a 30- to 50-ml screw-capped plastic centrifuge tube by centrifugation at 4000 g for 5 rain at room temperature (never cool). Wash the pellet at room temperature with 5 ml of 50 mM Tris-HCl, pH 7.5. Resuspend the pellet in 5 ml of 50 mM Tris-HCl, pH 8.5, and incubate the cells for 30 to 60 min at 37° with slow shaking. 15 W. Takahashi, H. Yamagata, K. Yamaguchi, N. Tsukagoshi, and S. Udaka, J. Bacteriol. 156, 1130 (1983).
30
VECTORS FOR EXPRESSING CLONED GENES
[3]
4. Spin as above and wash the cell pellet with 1 ml of MTP. Spin again and resuspend the pellet evenly in 0.5 ml of MTP. 5. Add plasmid DNA dissolved in less than 50/xl of TE or MTP to the cell suspension and mix well. Quickly, add 1.5 ml of the PEG solution and immediately mix well without vigorous agitation. Keep the mixture at room temperature for about 2 min with occasional gentle mixing. Add 5 ml of MTP and mix well. Collect the cells by centrifugation at 4000 g for 10 min at room temperature. Suspend them in 1 ml of T2U medium containing 20 mM MgC12 and incubate at 37° for 2.5 hr with moderate shaking. When erythromycin is used as a selective drug, it must be added to a final concentration of 0.1 /~g/ml after 30 min of incubation. 6. Spread aliquots (0.1-0.2 ml) of the culture on T2U agar plates containing the selective drug (I0/xg/ml of erythromycin or 60/xg/ml of neomycin) and incubate the plates at 30 or 37°. Colonies should appear after about 2 days at 37° or after about 3 days at 30 °. Transformation of B. brevis with plasmids harboring a heterologous gene is often more successful at 30°. Some 104 to 105 transformants can be obtained when 1/zg of intact pNU210 is used.
Electroporation 1. Take a small amount ofB. brevis 47-5Q cells grown on a T2U plate, inoculate it into 5 ml of T2U medium, and grow overnight with shaking at 37°. Dilute the overnight culture 100-fold in 100 ml of the same medium and incubate at 37° with vigorous shaking for approximately 4.5 hr. 2. At the early stationary phase of growth (the OD660is approximately 3.5), chill the culture in an ice/water bath, and then collect cells in a 500-ml screw-capped plastic centrifuge tube by centrifugation at 4000 g for 5 min at 4 °. Wash the pellet at 4° with 200 ml of cold solution A, which contains 93 g of sucrose and 150 g of glycerol in 1 liter of 0.1 mM sodium phosphate buffer, pH 7.4. Then wash the cells with 100 ml and then with 4 ml of cold solution A. Resuspend the washed cells in 0.5 ml of cold solution A. Transfer 45 /~1 each of the cell suspension to small plastic centrifuge tubes. Cool the tubes to - 7 0 ° in a dry ice-ethanol bath and stock them in a - 8 0 ° freezer (competent cells). 3. Take out the tube containing frozen competent cells and keep it cold. Add plasmid DNA dissolved in less than 2 /.d of TE to the cell suspension and mix. Transfer the cell suspension to a 0. l-cm cuvette of the Gene pulser apparatus (Bio-Rad, Richmond, CA). Set the apparatus at 0.9 kV, 25/xF, and 200 l). After delivering the pulse, quickly add 1 ml of cold T2U medium containing 20 mM MgCI2, keep it at 4 ° for 10 min, and then incubate at 37° with shaking for 2.5 hr. Spread aliquots of the
[3]
HETEROLOGOUS PROTEIN PRODUCTION BY B. brevis
31
culture on T2U agar plates containing the selective drug. Other details are similar to the procedure described above in the previous section. The frequency is about 105 transformants//xg of DNA. Salts in the plasmid DNA solution reduce the transformation frequency. Production of Heterologous Proteins Clones that produce heterologous proteins of bacterial origin are obtained rather easily by the procedure described above. On the other hand, transformants that efficiently produce mammalian proteins are found often at a low frequency (sometimes very low). Mammalian proteins are often toxic to B. brevis cells so that cells producing such a protein grow slowly or die rather easily. Therefore it is necessary to screen a large number of transformants for particular clones that are producing large amounts of mammalian protein. Unless a special procedure to detect the heterologous protein is available, immunoassays are convenient for measuring the protein productivity of clones. Plates with transformants are covered with a sterile membrane (e.g., pure nitrocellulose membrane; Bio-Rad) for 1-3 hr. After removal, the membrane is treated with antibody against the heterologous protein and processed for a color reaction using a standard protocol, t6 The area of the membrane corresponding to protein-producing colonies is stained. The size of the stained halo varies depending on the extent of protein production and the type of protein. Identified transformants that produce the protein are immediately picked up from plates and grown on fresh plates. After 1 day of growth at 37 ° (or 30°), cells are removed from the plates, resuspended in T2 medium plus 20% (w/w) glycerol, and stored at or below - 8 0 °. Several clones thus selected are examined by culturing in liquid medium to measure the amount of protein secreted. The production efficiency for each protein varies greatly with the culture conditions, which include the medium composition and growth temperature. Optimal conditions are unique for each protein and must be determined for efficient production. We found that varying the amounts of MgC12 and CaC12 in PM medium often allowed efficient protein production. Also, the addition of glucose to a final concentration of 3% (w/v) after 2 days of growth often improved the efficiency of protein production.
Examples of Secretory Protein Production We succeeded in producing a number of proteins by the B. brevis system. Most of the bacterial enzymes tested could be produced at a yield t6 V. Nagarajan, this series, Vol. 185, p. 214.
32
VECTORS FOR EXPRESSING CLONED GENES
[3]
of more than 1 g/liter. For example, about 3 g/liter thermophilic a-amylase of Bacillus stearothermophilus was secreted. 17On the other hand, production of mammalian proteins was often much less efficient than that of bacterial proteins, although it was much more efficient than production by other hosts such as B. subtilis, E. coli, and Saccharomyces cerevisiae. So far, among mammalian proteins, human epidermal growth factor (EGF) was the one most efficiently produced [0.24 g/liter ~8and 1 g/liter (unpublished observations). Active human salivary a-amylase 2 and swine pepsinogen 19 were secreted up to 60 and 11 mg/liter, respectively. High-level production of EGF and human a-amylase 2 was achieved only after extensive improvements were made mainly for the vector (use of a derivative of pHY481) and host (mutant isolation), respectively. Certain animal proteins such as human interleukin 2 were secreted in large amounts (more than 50 rag/liter) only when the signal peptide was altered to become more hydrophobic (e.g., leucine tripeptide was inserted into the hydrophobic region of the MWP signal peptide). Discussion of Problems As described above, transformants that produce mammalian proteins are often found at a low frequency, probably because efficient production of the proteins is toxic for bacterial cells. One way to circumvent this problem is to use B. subtilis as an initial host. In B. subtilis, cwp promoters direct only weak expression and hence the plasmid with the desired construct tends to be stably maintained. All B. brevis vectors described here can replicate in B. subtilis and the drug-resistance genes on the plasmids can be used as selective markers. A large amount of plasmid DNA can be prepared from B. subtilis transformants. This DNA can be used to transform B. brevis to obtain a large number of transformants. Even when positive transformants of B. brevis are found, these cells sometimes cannot maintain the correct plasmid. Deletions of the foreign gene or cwp promoter are frequently found. Serial single-colony isolation of positive clones is helpful. Mutagenesis ofB. brevis with N-methyl-N'nitro-N-nitrosoguanidine (NTG) prior to transformation is also helpful to obtain clones that can maintain the plasmid with the correct structure. When clones maintaining the plasmid can be obtained but the amount of foreign proteins produced is not large, mutants producing the protein 17 H. Takagi, A. Miyauchi, K. Kadowaki, and S. Udaka, Agric. Biol. Chem. 53, 2279 (1989). ~8 H. Yamagata, K. Nakahama, Y. Suzuki, A. Kakinuma, N. Tsukagoshi, and S. Udaka, Proc. Natl. Acad. Sci. U.S.A. 86, 3589 (1989). 19 M. Takao, T. Morioka, H. Yamagata, N. Tsukagoshi, and S. Udaka, Appl. MicrobioL Biotechnol. 30, 75 (1989).
[3]
HETEROLOGOUS PROTEIN PRODUCTION BY B. brevis
33
with improved yields can be isolated by mutagenesis of the clones with NTG. The procedures to mutagenize B. brevis with NTG are as follows. 1. Grow B. brevis freshly in 5 ml of T2 medium at 30° with shaking until the OD660 is 0.6. 2. Collect the cells by centrifugation for 5 min at 4000 g at room temperature and wash the cells with 5 ml of 200 mM phosphate buffer (KHzPO4-NaEHPO 4 , pH 6.4) and centrifuge again as above. 3. Resuspend the cells in 0.5 ml of phosphate buffer containing 200 t~g/ml of NTG and incubate the cell suspension at 30° for 30 min. 4. Wash the cells with 5 ml of phosphate buffer as above and resuspend the cells in 5 ml of T2 medium supplemented with 20 mM MgC12 . 5. Grow the cells for more than 3 hr at 30° with shaking. After a 102to 104-fold dilution with T2 medium, spread 0.1 ml of the cell suspension on an appropriate plate. Conclusion Taking advantage of the unique characteristics of B. brevis, which secretes large amounts of proteins into the medium but hardly any proteases, we have developed a novel host-vector system for efficient synthesis and secretion of foreign proteins. The multiple promoters and the signal peptide-coding region of the gene for one of the major cell wall proteins of B. brevis 47 were used to construct expression-secretion vectors. With this system, many bacterial proteins and human epidermal growth factor were efficiently secreted at yields of more than 1 g/liter. The yield of other mammalian proteins was less, but still l0 to 100 times higher than has been reported with other systems. In addition to the direct use of the produced proteins, this system should be useful for engineering proteins by random or localized mutagenesis. Because the active proteins are secreted efficiently into the medium, clones producing proteins of altered properties can be easily screened by direct assay of the culture medium.
34
VECTORS FOR EXPRESSING CLONED GENES
[4]
[4] U s e o f M u l t i d r u g R e s i s t a n c e G e n e in M a m m a l i a n Expression Vectors
By SUSAN E. KANE and MICHAEL M. GOTTESMAN Introduction As the fields of molecular biology, cell biology, and human genetics converge, and technological advancements in gene cloning, gene expression, and gene therapy multiply, the need for efficient, flexible systems for expressing foreign genes in mammalian cells is increasing. The terms "foreign" and "heterologous" genes will be used interchangeably in this chapter to refer to cloned DNAs of interest that are to be expressed in tissue culture cells or in animals. These DNAs generally do not have a selectable phenotype and must either be expressed transiently or cotransferred into cells with a selectable marker. For general reviews of gene expression and expression vectors, see Refs. 1 and 2. Virus-based vector systems, such as retroviruses, vaccinia virus, and bovine papiUomavirus, have proved useful for both stable and transient gene expression. Gene transfer can be efficient and reproducible with such viruses and infection of a variety of cell types is often possible. Many plasmid vectors make use of strong promoter elements for high-level gene expression in transfected cells, and tissue-specific or inducible promoters allow expression under more controlled conditions. Amplifiable selection systems have been developed for stable overexpression of foreign sequences, the most common of these being based on dihydrofolate reductase, which confers resistance to folate analogs. This chapter will discuss the theory and applications of another such amplifiable selection system, based on the human multidrug resistance gene (MDR1), and resistance to a variety of clinically relevant drugs, such as colchicine, adriamycin (doxorubicin) and vinblastine. Multidrug Resistance as Dominant Selectable Marker The discovery that cultured cells could develop cross-resistance to multiple cytotoxic drugs led to the cloning of the mouse 3 and human 4 mdr S e e a r t i c l e s in this s e r i e s , Vol. 185, pp. 4 8 5 - 6 1 1 . 2 S. E. K a n e , Genet. Eng. 13, 167 (1991). 3 p. G r o s , J. C r o o p , a n d D. H o u s m a n , Cell 47, 371 (1986).
4C.-J. Chen,J. E. Chin, K. Ueda, D. Clark, I. Pastan, M. M. Gottesman,and I. B. Roninson, Cell 47, 381 (1986). METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
[4]
M D R 1 GENE IN MAMMALIAN EXPRESSION VECTORS
35
cDNAs. These cDNAs encode a 170,000-Da transmembrane glycoprotein that is an energy-dependent drug efflux pump known variously as P170, P-glycoprotein, or the multidrug transporter. Increased expression of the multidrug transporter leads to resistance to a variety of cytotoxic drugs, including the anticancer drugs doxorubicin, daunorubicin, vinblastine, vincristine, VP-16, VM-26, actinomycin D, and taxol, and other cytotoxic agents such as colchicine, puromycin, emetine, ethidium bromide, and mithramycin. Selections of increasing stringency (i.e., stepwise increases in drug concentration) in any of these drugs results in cross-resistant cells with overexpression of the MDR1 gene (for human), frequently as a result of amplification of this gene (for a review, see Ref. 5). The cloning of the cDNA for the multidrug transporter made it possible to create retroviral expression vectors for the mouse 6 and human 7 cDNAs. When introduced into drug-sensitive cells, these vectors confer the complete phenotype of multidrug resistance on these cells. Virtually all cells, except those that are drug resistant to begin with, are susceptible to transformation with the MDR1 cDNA, making it a flexible, dominant selectable marker. An MDR1 cDNA was introduced into the germline of transgenic mice.S In one line of MDRl-transgenic mice, P-glycoprotein is expressed on the surface of bone marrow cells, rendering these mice resistant to the marrow-toxic effects of a variety of natural product anticancer drugs. 9 Thus, the MDR1 cDNA is a good dominant selectable marker in vivo as well as in vitro.
Applications of MDR1-Based Selections to Foreign Gene Expression The cloned MDR1 cDNA has been adapted for use in gene transfection of tissue culture cells, in retroviral transfer of foreign sequences, and in bone marrow transplantation studies. MDR1 expression can be selected by growth with a variety of cytotoxic drugs (see above) or by cell-sorting technology (see [14] this volume). Coexpression of heterologous coding sequences as well as antisense or catalytic RNA sequences can be achieved in a variety of cell types. 5 S. E. Kane, I. Pastan, and M. M. Gottesman, J. Bioenerg. Biomembr. 22, 593 (1990). 6 p. Gros, Y. B. Neriah, J. M. Croop, and D. E. Housman, Nature (London) 323, 728 (1986). 7 K. Ueda, C. Cardarelli, M. M. Gottesman, and I. Pastan, Proc. Natl. Acad. Sci. U.S.A. 84, 3004 (1987). 8 H. Galski, M. Sullivan, M. C. Willingham, K.-V. Chin, M. M. Gottesman, I. Pastan, and G. T. Merlino, Mol. Cell. Biol. 9, 4357 (1989). 9 G. Mickisch, G. T. Merlino, H. Galski, M. M. Gottesman, and I. Pastan, Proc. Natl. Acad. Sci. U.S.A. 88, 547 (1991).
36
VECTORS FOR EXPRESSING CLONED GENES
[4]
Cotransfection and Coamplification
The most common application of MDR1 selection is in tissue culture transfection experiments. MDR1 as a selectable marker was originally used in cotransfections, with MDR1 and foreign cDNAs carried on separate plasmids. 1°'11 The MDR1 plasmid, called pHaMDR, has the human MDR1 cDNA under the control of Harvey murine sarcoma virus long terminal repeats (LTRs). Foreign sequences on a separate plasmid can be regulated by any transcription control elements. Using transfection methods that lead to tandem integration of cotransfected sequences, selection for uptake and expression of MDR1 results in cell lines that also express the foreign sequences with good efficiency. Subsequent selection for amplified expression of MDR1 allows coamplification of foreign gene expression as well. A modification to pHaMDR places MDR1 and heterologous sequences on the same plasmid molecule.12 The modified vector, termed pSK 1.MDR, maintains MDR1 under control of retroviral LTRs and includes a simian virus 40 (SV40) promoter and polyadenylation signal plus a unique cloning site for insertion of heterologous sequences. This vector is efficient for coexpression of MDR1 and foreign sequences and should be useful with transfection methods that yield single-copy or multiple, unlinked integrations of the transferred sequences. The most commonly used selecting agent in transfection experiments is colchicine because it is effective and inexpensive. Colchicine disrupts microtubules and thus inhibits cell division. Cells that are not multidrug resistant become multinucleated and eventually die in the presence of colchicine, while those that take up and express MDR1 are resistant to the drug. By increasing the selective pressure (colchicine concentration) on these cells, they must express progressively more MDR1 gene product to remain drug resistant. This is accomplished either by bona fide gene amplification ~° or by enrichment for cells in a population that already express high levels of MDR1. lz By either mechanism, the end result is a cell line that is resistant to high concentrations of colchicine and that also expresses high levels of cotransferred foreign sequences. Retroviral Transfer o f MDR1
The MDR1 expression and selection system has also been adapted for use as a retroviral vector. In pHaMDR/A, MDR1 cDNA (lacking its own l0 S. E. Kane, B. R. Troen, S. Gal, K. Ueda, I. Pastan, and M. M. Gottesman, Mol. Cell. Biol. 8, 3316 (1988). 11 R. Konig, G. Ashwell, and J. A. Hanover, Proc. Natl. Acad. Sci. U.S.A. 86, 9188 (1989). 12 S. E. Kane, D. H. Reinhard, C. M. Fordis, I. Pastan, and M. M. Gottesman, Gene 84, 439 (1989).
[4]
M D R 1 GENE IN MAMMALIAN EXPRESSION VECTORS
37
polyadenylation signal) is flanked by retroviral LTRs and the primary MDR1 transcript contains viral packaging signals near the 5' end. The pHaMDR/A plasmid is transfected into retrovirus packaging cell lines, T M transfected cells are selected for resistance to colchicine, and virus containing the MDR1 coding sequences is isolated from the culture supernatant. This approach has been used to isolate virus with the human MDR1 gene ~5 and, in a separate construction, with a mouse mdr gene. ~6 These viruses can infect and confer multidrug resistance on rodent, dog, and human cell lines in culture. In addition, primary mouse bone marrow cells have been infected in vitro with M D R I virus and multidrug resistant granulocyte-macrophage progenitor colonies were subsequently isolated by colchicine selection. Bone marrow infected with the MDR1 virus can reconstitute mice and form spleen foci containing M D R ! cDNA sequences with a moderate efficiency.17 These results suggest that M D R ! can be used as a selectable marker in a retroviral vector system. To demonstrate that MDR1 virus can also be useful in cotransfer experiments, Germann et al. constructed a fusion gene, with M D R 1 and human adenosine deaminase (ADA) coding sequences linked to encode a fusion protein of the two gene products. 18,~9 When virus carrying the fusion gene is used to infect mouse fibroblasts, a bifunctional fusion protein is produced. Thus, infected cells selected for resistance to colchicine express MDR1 and also express functional ADA. ~9 The fusion protein is membrane associated in the drug-resistant cells, consistent with the membrane localization of the MDR1 gene product acting as a multidrug transporter. When transformed mouse cells are infected with the fusion gene virus, resulting drug-resistant cells can be used to form tumors in nude mice. Such tumors grow in vivo in the absence of drug and maintain expression of the bifunctional fusion protein even when removed from the mouse and grown in culture without colchicine. ~9 This work with MDR fusion proteins and retroviral gene transfer indicates that the MDR1 expression system has potential for use in mammalian gene therapy and bone marrow transplantation studies. It should also be possible to improve the retroviral vector either by encoding two distinct ~3 R. Mann, R. C. Mulligan, and D. Baltimore, Cell 33, 153 (1983). 14 A. D. Miller, M.-F. Law, and I. M, Verma, Mol. Cell. Biol. $, 431 (1984). 15 I. Pastan, M. M. Gottesman, K. Ueda, E. Lovelace, A. V. Rutherford, and M. C. Willingham, Proc. Natl. Acad. Sci. U.S.A. 85, 4486 (1988). 16 B. C. Guild, R. C. Mulligan, P. Gros, and D. E. Housman, Proc. Natl. Acad. Sci. U.S.A. 85, 1595 (1985). ~7 j. R. McLachlin, M. A. Eglitis, K. Ueda, P. W. Kantoff, I. H. Pastan, W. F. Anderson, and M. M. Gottesman, J. Natl. Cancer Inst. 82, 1260 (1990). ~8 U. A. Germann, M. M. Gottesman, and I. Pastan, J. Biol. Chem. 2,64, 7418 (1989). 19 U. A. Germann, K.-V. Chin, I. Pastan, and M. M. Gottesman, FASEB J. 4, 1501 (1990).
38
VECTORS FOR EXPRESSING CLONED GENES
[4]
genes, MDR1 and the heterologous gene, on a polycistronic messenger RNA or by including a separate promoter plus foreign sequences inside the LTRs. In either case, MDR1 and foreign genes should both be packaged into virions with some efficiency.
Selection and Amplification o f Multidrug-Resistant Cells Whether gene transfer is by DNA-mediated transfection or virus-mediated transduction, selection for multidrug resistance and subsequent amplification to high-level gene expression follow the same procedure. Although MDR1 can confer resistance to a variety of drugs, colchicine has been used exclusively in the studies discussed in this chapter. There is some evidence to suggest that the particular MDR1 allele included in the pHaMDR and pSK1.MDR vectors confers preferential resistance to colchicine over other drugs, 2° suggesting that selection with colchicine might be more efficient and yield more resistant colonies than selection with other agents. This possibility has not been tested. The selection and amplification processes are relatively simple and rapid. In general, transfected or infected cells are initially plated in a low concentration of colchicine (see p. 40 for details on drug concentrations) for 10-14 days. This allows cells that take up, stably integrate, and express MDR1 to grow into individual colonies; colonies are pooled or single colonies are expanded at the initial low drug concentration. Selection for amplified expression proceeds by stepwise increases in the concentration of colchicine in the culture medium. Cells expressing MDR1 gene product on the cell surface can also be selected by magnetic activated cell sorting (MACS) technology, using a monoclonal antibody directed against an external epitope of the multidrug transporter (see Ref. 21 and [14] this volume). While this approach is promising for sorting out drug-resistant cells from a mixed population of resistant and sensitive tumor cells, for example, MACS has not been adequately tested for performing initial selection of transfected cells that take up and express MDR1. A potential problem is that only a small percentage of cells that express P-glycoprotein on their cell surface at early times after transfection will go on to establish stable expression of the protein at longer times after transfection. Therefore, cells selected early with MACS might lose their transfected sequences with subsequent passage. Later selection with MACS might miss the small percentage of MDRl-expressing cells that have been overgrown by nonexpressing cells. 20 K. Choi, C.-J. Chen, M. Kriegler, and 1. B. Roninson, Cell 53, 519 (1988). 21 R. Padmanabhan, T. Tsuruo, S. Kane, M. Willingham, B. Howard, M. M. Gottesman, and I. Pastan, J. Natl. Cancer Inst. 83, 565 (1991).
[4]
M D R I GENE IN MAMMALIAN EXPRESSION VECTORS
39
The value of MACS or the related technique of fluorescence-activated cell sorting might be in sorting out very high MDR1 expressers in a population of cells previously selected for multidrug resistance. Cells, DNA, and Transfection Methods The MDRI expression system has been used in a variety of cell types. Initially tested in mouse fibroblasts, N I H 3T3 cells, 1° many human cells, such as epithelial A431 cells, 2~ KB-3-I fibroblasts, and SK-Hep hepatoma cells, as well as Chinese hamster ovary cells ~1 and transformed NIH 3T3 cells, ~° have been transfected with the MDR1 cDNA, The retrovirus carrying the MDR1 cDNA has been used to infect mouse bone marrow, 17 mouse myoblasts, 22 dog MDCK cells, 15 transformed NIH 3T3 cells, 19 as well as mouse leukemia and lymphoma cells, human FEM-X melanoma cells, human colon cancer HT-29 cells, and human breast cancer MCF-7 cells. It should also be possible to use other cell lines that are not inherently resistant to colchicine, because MDR1 is a dominant selectable marker. Cell type often determines the transfection method that must be utilized for DNA transfer (transfection techniques are reviewed in Ref. 1). The MDR1 system has been used primarily with calcium phosphate-mediated transfection and retroviral infection, as discussed above. The former method is particularly useful for introducing multiple tandem copies of the transferred sequences into the host cell genome. This might be important for establishing colonies that survive initial selection with colchicine. Once high-level expression is achieved by colchicine amplification, the integrated sequences appear to be stable on long-term culturing in the presence of drug. Other transfection methods, such as electroporation and lipofection, have not been rigorously tested yet with the MDRI system. It is possible that techniques that lead to single or low-copy integration events within a single cell will be less efficient at producing multidrug-resistant colonies, or that those colonies that do survive initial selection will not be amplified with good efficiency. However, the results with retroviral-mediated MDR1 transfer suggest that initial selection, at least, is possible with low MDR1 copy number. The discussion so far has focused on high-level expression of foreign DNAs that code for proteins. MDR1 selection and amplification have also been exploited for overexpression ofantisense RNA sequences (E. Cantin, personal communication), which are often required in high dosage relative to the level of their target mRNA. The MDR1 system should be adaptable 22 A. Salminen, H. F. Elson, L. A. Mickley, A. T. Fojo, and M. M. Gottesman, Hum. Gene Ther. 2, 15 (1991).
40
VECTORS FOR EXPRESSING CLONED GENES
[4]
for catalytic ribozyme expression as well, and for generating high-copy DNA or RNA elements that potentially interact with soluble trans-acting factors within cells. Methods Following is a description of DNA transfer, colchicine selection and amplification, and analysis of multidrug-resistant cells for expression of transferred sequences. The protocols primarily refer to using the pSK1.MDR expression vector with N I H 3T3 cells and calcium phosphatemediated transfection, with comments on variations included when relevant.
Cloning Foreign Sequences into pSK1.MDR Figure 1 is a partial restriction map of the pSK1.MDR vector. Using standard enzyme digestions and cloning techniques, foreign sequences are inserted into the unique SalI site [base pair (bp) 69 on the map] located between the SV40 promoter and polyadenylation signal. The orientation of the insert can be determined by restriction enzyme analysis, using the indicated restriction sites as reference. Transcription proceeds in a clockwise direction from the SV40 promoter, so insert orientation should be 5' to 3' (left to right) on this map. MDR1 transcription also proceeds clockwise from the LTR and can either terminate at its own polyadenylation signal or continue to the downstream LTR polyadenylation site.
Determination of Colchicine Dose Before beginning any transfection or infection with MDR1, it is important to determine the concentration of colchicine that will be used for the initial selection step. This can vary significantly from cell line to cell line. With N I H 3T3 cells, transfectants are selected with 60 ng/ml ofcolchicine. Most human cell lines tend to be 10-fold more sensitive to colchicine, however, and should be selected with approximately 6 ng/ml of colchicine. Colchicine inhibition of human cells is sensitive to drug concentration, however, so the exact dosage to use should be determined carefully. At the other end of the spectrum, Chinese hamster ovary cells are inherently resistant to colchicine and require concentrations of 200 ng/ml or more of drug. To determine the concentration of colchicine for selection: 1. In standard growth medium for the cells, seed 100-300 cells into each well of a 12-well tissue culture plate. Allow the cells to attach to the bottom of the well.
[4]
MDRI GENE IN MAMMALIANEXPRESSIONVECTORS
41
Sphl 15453I Sphl 15381 I
BamHI 15116. /
44 69
SV40
Hindlll Pstl 1
2852
Pstl 12171 3635
pSK1.MDR 15585 bp
Stul 11189 J Hindlll 11147
LTR
Pstl 10122 J Stul 9822
~ ~ 1 BamHI 6031 BamHI 6062
Stul 8916 I Hindlll 8791 FIG. 1. A partial restriction map of the pSKI .MDR expression vector. Foreign DNAs are inserted into the unique SalI restriction site at bp 69. Expression of the foreign sequences is controlled by an SV40 promoter and polyadenylation signal (black box). Expression of the MDR1 selectable marker is controlled by retroviral LTRs (gray boxes). Arrows indicate direction of transcription. Included on the map are all the sites recognized by restriction enzymes BamHI, HindlII, PstI, SphI, and StuI. These enzymes are useful for determining the orientation of foreign sequences inserted at the Sa/I site.
2. R e m o v e the m e d i u m and replace with complete medium plus increasing concentrations o f colchicine. F o r example: 0, 2, 3, 4, 5, 6, 7, and 8 ng/ml of colchicine in duplicate wells for h u m a n cells; 0, 10, 20, 30, 40, 50, 60, 70, 80, and 100 ng/ml in duplicate wells for mouse cells. 3. I n c u b a t e cells under standard t e m p e r a t u r e and CO2 conditions for the cells, for 5-10 days. 4. R e m o v e the m e d i u m and stain cells with 0.5% (v/v) methylene blue in 50% (v/v) ethanol.
42
VECTORS FOR EXPRESSING CLONED GENES
[4]
5. Choose the concentration of drug at which no cells grow to form colonies. Confirm that this concentration of drug gives no (or few) surviving colonies when cells are plated at transfection densities (5 x 105 cells/ 100-mm tissue culture dish).
Transfection and Initial Selection Calcium phosphate-mediated transfections are performed essentially as described by Graham and van der E b 23 and Gorman et al. 24
Reagents CaCI 2, 2.5 M Sterile HzO HBS (2 x): 50 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), pH 7.1,280 mM NaCI, 1.5 mM Na:HPO4 Filter sterilize all reagents and store at 4 °. The pH of the 2 x HBS is critical and should be checked periodically and adjusted to pH 7.05-7.1 if necessary.
Protocol l. On the afternoon of day 0, plate 5 x 105 cells/10-cm tissue culture dish. Set up one dish per DNA to be transfected plus a negative control dish for a mock transfection without DNA. 2. On day 1, change the medium of the cells 3 hr before the transfection. Use complete medium. 3. For the transfection, set up two 15-ml conical tubes for each transfection: tube I: 250/zl 2 x HBS tube 2: D N A , 24a 25 ~1 2.5 M CaC12 , H20 to 250/zl 4. Using a 1-ml pipette, bubble air into the solution of 2 x HBS; at the same time, add DNA solution dropwise to the HBS. Continue bubbling air through the solution for 5 sec after all the DNA is added. 5. Let the mixture sit at room temperature for 30 min to allow precipitate to form. 6. Add DNA/HBS mix to cells in complete medium, dropwise with
23 F. L. Graham and A. J. van der Eb, J. Virol. 52, 456 (1973). 24 C. M. Gorman, G. T. Merlino, M. C. Willingham, I. Pastan, and B. H. Howard, Proc. Natl. Acad. Sci. U.S.A. 79, 6777 (1982). 24a Use a total of 5-10/~g DNA. If using pSK1.MDR, 5-10/zg; if using pHaMDR plus a separate plasmid with foreign sequences, use approximately equal molar amounts of the two plasmids.
[4]
MDR1 GENE IN MAMMALIAN EXPRESSION VECTORS
43
gentle swirling. For stable transfections, allow the precipitate to remain on the cells overnight ( - 1 6 hr). 7. On day 2, wash the cells two or three times with phosphate-buffered saline (PBS) (plus Ca2+/Mg 2+) or serum-free medium and add fresh complete medium overnight. 8. On day 3, split the cells 1 : 4 into selective medium, 60 ng/ml colchicine for NIH 3T3 cells. Change the medium once, after 5 days in colchicine, being careful not to disrupt colonies. Notes 1. On day 3, the cells should be split into three or four dishes or flasks, depending on the purpose of the experiment. In general, include one dish for staining colonies to quantitate transfection efficiency and the rest for picking individual colonies or for pooling colonies to expand and amplify in colchicine (see below). The density at which cells should be plated at this stage will vary from cell type to cell type. Under the conditions described above, N I H 3T3 cells split 1 : 4 on day 3 will result in 100-300 colonies per 10-cm dish by 10-14 days after addition of colchicine. 2. Colchicine is stored as 10- or 1-mg/ml stock solutions, in dimethyl sulfoxide (DMSO). Tubes are kept wrapped in foil and frozen at - 2 0 °. The concentration of DMSO in selection medium should not exceed 0.5% (v/v). 3. Colonies will appear by 7-14 days after selection is initiated (again, this should vary with cell type). To stain colonies, remove the growth medium and replace with 5 ml 0.5% (w/v) methylene blue in 50% (v/v) ethanol. Let sit for 5 min on the cells and wash off with several rinses of H20. Background will appear light blue because drug-sensitive cells are not killed by colchicine but are prevented from dividing. Infection of Drug-Sensitive Cells with an MDR1 Retrovirus The pHaMDRA vector has been packaged in both ecotropic and amphotropic mouse 3T3 packaging cell lines.~5 Viral titers of approximately 8 x 104 transforming units per milliliter of supernatant have been obtained. To infect drug-sensitive cell lines with these supernatants we use the following protocol: 1. Cells are plated at 1-2 x 104 cells/100-mm dish and allowed to attach to the plastic substrate overnight. Up to 105 cells/dish can be used if retroviral transformation frequencies are low. 2. Viral supernatant at a dilution of 1/10 or 1/5 (1 or 2 ml of virus in a total of 10 ml of complete medium containing 2/zg/ml Polybrene, Abbott Laboratories) is added to the cell monolayer. Higher dilutions of virus can also be used.
44
VECTORS FOR EXPRESSING CLONED GENES
[4]
3. After 24 hr, virus is removed and selection medium containing an appropriate concentration of colchicine is added (see above). 4. Colchicine-resistant colonies will appear after 7-14 days. A mocktreated control should show no colonies if the selection conditions are appropriate.
Amplification in Colchicine Before beginning the amplification process, it is important to have healthy, drug-resistant cells growing at a low drug concentration (60-80 ng/ml of colchicine for N I H 3T3 cells). Individual colonies can be isolated and expanded until they are growing in T75s. Alternatively, a dish of selected colonies can be trypsinized and pooled into a single T75 flask and grown for two or three passages at low drug concentration until only drug-resistant cells persist. Some multinucleated drug-sensitive cells might survive the early passages.
Protocol 1. From a confluent T75 or 10-cm dish at low drug concentration, trypsinize cells and replate in a twofold higher drug concentration. Do not plate cells at less than about 25% confluence at any time during the amplification process. 2. Maintain cells at this drug concentration for at least two passages. 3. Repeat the amplification in twofold increments of drug concentration, again passaging at least two times at each level and maintaining at least 25% confluence.
Notes 1. It is a good idea to freeze aliquots of cells at each drug concentration during amplification, and maintain some cells at selected concentrations for further analysis (see the next section). 2. The maximum drug concentration to which cells can be amplified will vary. With N I H 3T3 cells, transferred DNA is stable through longterm culturing in 1/.tg/ml colchicine. 3'5 Chinese hamster ovary cell amplification has been reported to 12.8/xg/ml colchicine, n Analysis of Multidrug-Resistant Cells Once individual clones or populations of cells have been amplified to high colchicine resistance, they can be analyzed by a variety of methods to confirm that they harbor and express human MDR1 as well as the foreign sequences of interest. Described briefly below are some approaches and tools for measuring MDR1 expression.
[4]
M D R I GENE IN MAMMALIAN EXPRESSION VECTORS
45
DNA Detection. Transfected MDR1 DNA can be detected by isolating genomic DNA from multidrug-resistant cells and performing Southern analysis 25 on EcoRI restriction enzyme digestion products. Filters are hybridized with a 3-kb EcoRI fragment derived from human MDR1 cDNA 26that detects a 3-kb band specific to transfected MDR1 sequences. This probe is completely internal to the MDR! cDNA, so it will not detect the number of different integration sites per cell line. To determine whether the transferred sequences are truly amplified during the course of stepwise colchicine selection, individual clones must be isolated, expanded, and selected in increasing concentrations of colchicine. Southern analysis should be performed on genomic DNA from cells growing at each drug concentration. The intensity of the 3-kb MDR1 band should increase as multidrug resistance and amplification increase. RNA Detection. MDR1 RNA can be measured by Northern analysis 27 or by primer extension. 28 In both cases, probes are available that will distinguish between transfected MDRl-specific transcription and transcription from endogenous genes in the transfected cell lines. Using human cells as transfection recipients, transfected MDRI and host MDRI detected by the probes are different in size. With nonhuman cells as host, the probes do not cross-react with other species of the multidrug resistance mRNA. Using the same 3-kb EcoRI fragment mentioned above as probe in a Northern hybridization, MDR1 mRNA transcribed from pSK1.MDR or from pHaMDR will be detected as 6- and 11-kb species. 7 Endogenous human MDR1 mRNA is 4.5 kb. Primer extension analysis using a human MDRl-specific primer 26 results in transfected and endogenous extension products that are different in size by about 13 nucleotides. As for DNA, RNA analysis should be performed on individual amplified clones of transfected cells rather than on mixed populations of cells in order to measure amplification of expression during the course of colchicine selection. Analysis for P-Glycoprotein Immunoprecipitation, Western blotting, indirect immunofluorescence, and fluorescence-activated cell analysis (FACS) can all be used to measure expression of P170, the MDR1 gene product. With the MDR1 selection 25 E. M. Southern, J. Biol. Chem. 98, 503 (1975). ,.6 K. Ueda, D, P. Clark, C.-J. Chen, I. B. Roninson, M. M. Gottesman, and I. Pastan, J. Biol. Chem. 262, 505 (1987). ,~7p. S. Thomas, Proc. Natl, Acad. Sci. U.S.A. 77, 5201 (1980). ~8 G. Merlino, J. S. Tyagi, B. de Crombrugghe, and I. Pastan, J. Biol. Chem. 257, 7254 (1982).
46
VECTORS FOR EXPRESSING CLONED GENES
[4]
system described in this chapter, the correlation between high MDR1 expression and high expression of cotransfected sequences is excellent (see Refs. 3 and 5; S. Kane and C. M. Fordis, unpublished observations). In addition, expression can be very stable. As stated earlier, amplified cultures can be maintained in colchicine for months without loss of MDR1 or cotransfected sequences. Furthermore, in transfected NIH 3T3 cells, amplified cells can be grown in the absence of colchicine for at least 2 weeks and still maintain expression of cotransfected sequences (S. Kane, unpublished observations). The best reagent for immunofluorescence and FACS is a monoclonal antibody called MRK-16. 29 MRK-16 is directed against an epitope of human MDRI that lies on the outside of the cell when P170 is in its native conformation in the plasma membrane of multidrug-resistant cells. Immunofluorescence can be used to analyze a population of cells for variability in MDR1 expression or an individual clone for level of expression. It can also be used preparatively by staining live colonies to identify clones that express significant levels of P170. These colonies can be isolated and expanded if they are already growing in high colchicine concentration, or they can be further amplified if they are at lower concentrations of drug. Fluorescence-activated cell analysis can likewise be used to identify high MDR1 expressers and cell-sorting capability can be used preparatively to isolate high expressers from a mixed population. As mentioned previously, magnetic bead-sorting technology, using the MRK-16 antibody, has been applied to MDR1 cell sorting and is described elsewhere in this volume. Immunoprecipitations can be performed by standard methods, with the following special considerations. 1. P170 is a stable protein (half-life, 48-72 hr), requiring long labeling times to detect newly synthesized protein. Label cells with [35S]methionine or a comparable labeling-grade mix of 35S-labeled amino acids, 200 /zCi/ml of labeling medium. Label overnight in 5 ml of culture medium (lacking methionine) plus 5% (v/v) serum. 2. The MRK-16 monoclonal antibody or C219 (Centocor, Malvern, PA) monoclonal antibody can be used in immunoprecipitations. In addition, polyclonal antibody 400730 is highly efficient in immunoprecipitations. Polyclonal antibody 4007 will cross-react with nonhuman species of P170,
29 H. Hamada and T. Tsuruo, Proc. Natl. Acad. Sci. U.S.A. 83, 7785 (1986). 30 S. Tanaka, S. J. Currier, E. P. Bruggemann, K. Ueda, U. A. Germann, I. Pastan, and M. M. Gottesman, Biochem. Biophys. Res. Commun. 166, 180 (1990).
[5]
HIGH-LEVEL GENE EXPRESSION BY T7 RNA POLYMERASE
47
and C219 is not entirely specific for P-glycoprotein.31 A series of antibodies directed against human P170 peptides has been described. 32Most of these antibodies, except for MRK-16, can be used on Western blots. 3. After resuspending immunoprecipitated material in sodium dodecylsulfate (SDS)-sample buffer for a Laemmli-type gel,33 elute antibody-P170 conjugate at room temperature for 15 min. Boiling will cause P170 to aggregate and it will not enter an SDS-polyacrylamide (PAGE) gel. 4. The MDRI gene product will migrate on SDS-PAGE as a diffuse band of approximately 170K. Apparent molecular weight will vary depending on the cell type being used for expression, probably due to differential glycosylation patterns. Cell lines expressing P-glycoprotein on their surfaces can also be detected by indirect immunofluorescence as described by Willingham et al.34 Positive cells can also be detected and sorted by FACS. For this purpose, we use 10/xg MRK-16/106 cells 29 with 85 /zg fluorescein isothiocyanate (FITC) goat anti-mouse IgG per 10 6 cells as a second antibody for fluorescence detection. 3~ F. Thiebaut, T. Tsuruo, H. Hamada, M. M. Gottesman, I. Pastan, and M. C. Willingham. J, Histochem. Cytochem. 37, 159 (1989). 3: E. P. Bruggemann, V. Chaudhary, M. M. Gottesman, and I. Pastan, BioTechniques 10, 202 (1991). 33 U. K. Laemmli, Nature (London) 227, 680 (1970). 34 M. C. Willingharn, N. D. Richert, M. M. Cornwel|, T. Tsuruo, H. Hamada, M. M. Gottesman, and I. Pastan, J. Histochem. Cytochem. 35, 1451 (1987).
[5] S t a b l e H i g h - L e v e l G e n e E x p r e s s i o n in M a m m a l i a n by T7 Phage RNA Polymerase
Cells
By A N D R E LIEBER, VOLKER SANDIG, WOLFGANG SOMMER, SILVIA BAHRING, and MICHAEL STRAUSS Introduction Various routes for high-level foreign gene expression in mammalian cells have been explored over the years) '2 However, there are not that many different principles to choose from; the choice is between transient and stable expression systems. Expression vectors for the first kind of I M. Strauss, U. Kiessling, and M. Platzer, Biol. Zentralbl. 105, 209 (1986). 2 j. Sambrook and M.-J. Gething, Focus 10, 41 (1987).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
48
VECTORS FOR EXPRESSING CLONED GENES
[5]
systems are mainly derived from viruses, such as vaccinia virus, 3-5 baculovirus, 6'7 simian virus 40 (SV40), 8'9 and others. Some of these systems
allow for very high expression levels, but only for the short period of 1-3 days.2,5,7 High-level transient expression is therefore useful for producing large amounts of a particular protein for further biochemical studies. However, under certain demanding conditions, it is desirable to express foreign genes continuously at a relatively high level. Stable expression of foreign genes can be achieved using constructs that either integrate into the host cell genome or have the ability to replicate at a moderate level. Vectors of the second type are preferentially derived from bovine papilloma virus (BPV) 1°-12 and from Epstein-Barr virus (EBV). 13 Integrating plasmid vectors can be amplified by linkage to genes such as that for dihydrofolate reductase (DHFR) 14A5or metallothionein 16 and subsequent selection with increasing drug concentrations. However, when genes are amplified by more than 100-fold the increase in the expression level is often less than 10-fold. 1~Additionally, differences in the level of expression achieved with various strong viral or mammalian promoters are within one order of magnitude when analyzed in transient expression assays.~7 Differences in the expression levels between individual clones derived from one and the same transfection can be enormous, ranging between only a few nanograms and several micrograms per 106 cells/ml/day. N,15 In searching for a stable expression system that is not restricted by the availability of host cell transcription factors and RNA polymerase, we became interested in the use of a bacteriophage promoter/polymerase system for which a highly selective function in prokaryotic host cells was 3 D. Panicalli and E, Paoletti, Proc. Natl. Acad. Sci. U.S.A. 79, 4927 (1982). 4 M. Mackett, G. L. Smith, and B. Moss, Proc. Natl. Acad. Sci. U.S.A. 79, 7415 (1982). 5 T. R. Fuerst and B. Moss, J. Mol. Biol. 206, 333 (1989). 6 G. E. Smith, M. J. Fraser, and M. D. Summers, J. Virol. 46, 584 (1983). 7 V. A. Luckow and M. D. Summers, Bio/Technology 6, 47 (1988). 8 M.-J. Gething and J. Sambrook, Nature (London) 293, 620 (1981). 9 j. T. Elder, R. A. Spritz, and S. M. Weissman, Annu. Rev. Genet. 15, 295 (1981). 10N. Sarver, P. Gruss, M.-F. Law, G. Khoury, and P. M. Howley, Mol. Cell. Biol, 1, 486 (1981). ii N. Hsiung, R. Fitts, S. Wilson, A. Milne, and D. Hamer, J. Mol. Appl. Genet. 2, 497 (1984). 12 D. DiMaio, in "The Papillomaviruses," p. 293. Plenum, New York, 1987. 13 p. B. G. M. Belt, H. Groeneveld, W. J. Teubel, P. van de Putte, and C. Backendorf, Gene 84, 407 (1989). 14 R. J. Kaufman and P. A. Sharp, J. Mol. Biol. 159, 601 (1982). 15 R. J. Kaufman, L. C. Wasley, A. J. Spiliotes, S. D. Gossels, S. A. Latt, G. R. Larsen, and R. M. Kay, Mol. Cell. Biol. 5, 1750 (1985). 16 G. N. Pavlakis and D. Hamer, Proc. Natl. Acad. Sci. U.S.A. 80, 397 (1983). 17 D. R. Hurwitz, R. Hodges, W. Drohan, and N. Sarver, Nucleic Acids Res. 15, 7137 (1987).
[5]
HIGH-LEVEL GENE EXPRESSION BY
T7 R N A POLYMERASE
49
known. Like other uneven T phages, the T7 phage codes for an RNA polymerase that is selective for phage promoters.~8-2° The latter, in turn, are so different in sequence from bacterial promoters that they cannot be transcribed by the bacterial RNA polymerase. ~8Moss and co-workers 2~'20were first in using this principle for the short-term vaccinia expression system. In this chapter we describe the adaptation of the transcriptional machinery of phage T7 to the mammalian cell system for highly selective stable gene expression. Principle In our original publication of T7 RNA polymerase-dependent gene expression in mammalian cells 23 we showed that the expression of the T7 RNA polymerase gene leads to cytoplasmic localization of the polymerase. This, in fact, is potentially useful for efficient transient expression from transfected plasmids and it is an ideal prerequisite for expression using vaccinia vectors. 24 For stable foreign gene expression a nuclear localization of the T7 polymerase is required. This has been achieved by fusion of a nuclear location signal to the amino terminus of the polymerase 25 or by substituting the nuclear location signal for the N-terminal 5% of the polymerase. 23 Cell lines can be established with this modified gene that harbor the T7 polymerase exclusively in the nucleus. Transfection of a foreign gene under control of a T7 promoter into such cells should result in transient as well as stable expression of the particular gene driven by the T7 polymerase. In fact, using chloramphenicol acetyl transferase (cat) 23 and other reporter genes we could demonstrate efficient expression within the range obtained with the strong Rous sarcoma virus (RSV) and cytomegalovirus (CMV) promoters. Human growth hormone has been expressed stably at levels of 20-30/zg/ml/106 cells/day. 26 However, we verified an effect that has been noticed before by ourselves and by other 18 j. j. Dunn and F. W. Studier, J. Mol. Biol. 166, 477 (1983). 19 p. Davanloo, A. H. Rosenberg, J. J. Dunn, and F. W. Studier, Proc. Natl. Acad. Sci. U.S.A. 81, 2035 (1984). 20 B. A. Moffat, J. J. Dunn, and F. W. Studier, J. Mol. Biol. 173, 265 (1984). 21 T. R. Fuerst, E. G. Niles, F. W. Studier, and B. Moss, Proc. Natl. Acad. Sci. U.S.A. 83, 8122 (1986). 22 B. Moss, O. Elroy-Stein, T. Mizukami, W. A. Alexander, and T. R. Fuerst, Nature (London) 348, 91 (1990). 23 A. Lieber, U. Kiessling, and M. Strauss, Nucleic Acids Res. 17, 8485 (1989). 24 O. Elroy-Stein and B. Moss, Proc. Natl. Acad. Sci. U,S.A. 87, 6743 (1990). 25 j. j. Dunn, B. Krippl, K. E. Bernstein, H. Westphal, and F. W. Studier, Gene 68, 259 (1988). 26 A. Lieber and M. Strauss, unpublished observations (1990).
50
VECTORS FOR EXPRESSING CLONED GENES
[5]
users of the system. It turned out that the T7 promoter can be transcribed by a cellular polymerase (probably polymerase II) quite efficiently if present in the context of a pGEM plasmid. This background expression level amounts to 10-50%, depending on the cell type. Thus, the T7 promoter in pGEM is actually a reasonably efficient promoter in mammalian cells, which allows the use of one and the same construct for expression in prokaryotic and eukaryotic cells as well as in a cell-free system. We will describe this interesting aspect elsewhere. 27 To establish an exclusive expression system for mammalian cells it was necessary to generate a modified T7 promoter that does not function with the cellular RNA polymerases and is still active with the T7 polymerase. Because we have succeeded in isolating several mutant promoters fulfilling these requirements we will describe the procedure in the first part of the experimental section (cloning and isolation of mutant T7 promoters). The strategy involves the synthesis of randomly mutagenized promoter oligonucleotides, their repeated incubation with HeLa cell nuclear extracts, the incubation of the unbound fraction with purified T7 polymerase, and amplification and cloning of the polymerase-bound fraction. Individual clones are assayed for function with T7 polymerase in vitro and within the cell using human growth hormone (hGH) as the test gene. The second part of the method to be outlined concerns gene transfer and the selection of clones expressing foreign genes under control of modified T7 promoters. The preferred protocol involves the use of a selectable marker also under control of a T7 promoter. There are three general variants of the procedure: (1) transfer of the gene of interest into cells already expressing the polymerase gene, (2) cotransfer of test gene, polymerase gene, and selectable marker, and (3) establishment of cells carrying a silent T7 promoter/test gene construct with subsequent gene activation by retroviral transfer of the polymerase gene. All three variants have their advantages and disadvantages. Finally, we will describe applications of the T7 expression system for purposes where this system is unique. Besides high-level expression of proteins from cDNA or genomic DNA it is exceptionally useful for expressing polycistronic mRNA and antisense RNA. Materials
Cell Lines
Mouse L t k - cells, Chinese hamster ovary (CHO) cells, and mouse myeloma cell line Sp2/0 Ag 14 are used as recipient cells for the polymerase gene and for cotransfections. 27 V. Sandig, A. Lieber, S. B~ihring, and M. Strauss, Gene, submitted (1992).
[5]
HIGH-LEVEL GENE EXPRESSION BY T 7 R N A POLYMERASE ~
I
1
f
r "Ec° R1369
\
I
51
f
-
/
I"7R N A !~ymemse ~
II T7 P.NA ! ~
~n~
1 ¢,, BsmHl --_ ~,_ ~ . . . ~ T R • Pat 11019
i LTR1 Pvu
M6SVTTN ~ 1 bp I12392 a 12467
rCLS/•x.bBtt
I
tit Pol
I"7R N A ~ y n ~ r a ~
FIG. 1. M a p s of plasmid v e c t o r s carrying the gene for a nuclear T7 R N A polymerase.
PMN clone A5 is a derivative of Ltk- cells expressing T7 RNA polymerase at a high level from the construct pMTT7N (Fig. 1) and has been selected with cadmium for the presence of the mouse metallothionein gene. 23 CHO185 is a clone of CHO cells that has been selected for the presence of pMTT7N. Ltk- cells, CHO cells, and their derivatives are grown in Dulbecco's modified Eagle's medium (DMEM) with 5% (v/v) fetal calf serum (GIBCO, Grand Island, NY) at 5% COz. Myeloma cells are grown in RPMI 1640 medium with 10% (v/v) fetal calf serum and 1 mg/ml of gentamycin.
Vectors and Oligonucleotides The vector family pGEM (Promega Madison, WI) was used originally for cloning and expression of reporter and selectable marker genes. In the studies described here only pT7neo is used from this series. For cloning
VECTORSFOR EXPRESSINGCLONEDGENES
52
HindI11.1~LHine!1.XImi ~ . _ ~_~Bam/~Bgm74
j
IS]
modit-a~l"l'7/a,omo~ !'''~Hindll125
"\
#
F,r.oR1396 BarnHI414 ~ / ~Xho!445 modit'l~lT7~ ~ ~ S p h 1445
hl~
7881~H~
~H
poly(A)signal Sph12640
Sph17.320
FIG. 2. Maps of plasmid vectors with marker genes driven from wild-type or mutant T7 promoters.
of the modified (mutant) T7 promoters pUC 19 is used. The standard vector with the most suitable promoter mutation (No. 86) is named pT7 m. Reporter and test genes are cloned into this vector and named accordingly pT7mhGH, pT7mluc, etc. (Fig. 2). The sequence of the synthetic T7 promoter is as follows: Wild type: Mutant 86:
5'-TTAATACGACTCACTATAGGGAGATA-3' ..... T ............ C .......
pMTT7N is derived from pMTT7 by replacing the BgllI/NarI fragment with a synthetic sequence coding for the nuclear location signal of SV40 large T antigen (Fig. 1). Because the sequence given in the original article 23 mistakenly contained two additional nucleotides, we give the correct sequence here (bold: sequence of the SV40 nuclear location signal):
[5]
HIGH-LEVEL GENE EXPRESSION BY T 7 R N A POLYMERASE
53
5'-AGATCTTTGCAAAAAGC T TTGC AAG A TGGATAAAGTT T TTAGAAAC TCCAGTAGG Me t A s pLy s V a 1 P h e A r g A s n S e r S e t A r g ACT
CCT CCA
A A A AAG AAG A G A AAG G T A GAA C G T C T A G A T C - 3 '
Thr-Pro-Pro-Ly
s -Ly s-Ly
s -Ar g-Ly
s -Va I -GI u ArgLeuAsp
pCMVT7N is constructed by inserting the 3.3-kbp BgllI/PvulI fragment of pMTT7N between the BgllI and SmaI sites of pCMVLT2. 28 pM6SVT7N contains the BgllI/PvulI fragment of pMTT7N inserted into the BgllI site of the retroviral vector construct pM6pac 29 (Fig. I).
Reagents All reagents should be of the highest purity available. The main compounds and enzymes used in our laboratory are from the following suppliers: Geneticin: GIBCO/BRL (Grand Island, NY)/Bethesda Research Laboratories (Gaithersburg, MD), Life Sciences Puromycin: Sigma (St. Louis, MO) DEAE-dextran (Mr 500,000): Pharmacia (Uppsala, Sweden) T7 RNA polymerase: New England BioLabs (Beverly, MA) Restriction enzymes: Boehringer Mannheim (Mannheim, Germany), Bethesda Research Laboratories, New England BioLabs Salts: Sigma BD-cellulose: Serva Feinchemica (Heidelberg, Germany)
Equipment Oligonucleotide synthesis: Performed using an Applied Biosystems (Foster City, CA) DNA synthesizer model 380 on the basis of methoxyamidites Electroporation experiments: Performed using the GenePulser (BioRad, Richmond, CA) Luciferase assays: Performed using a liquid scintillation counter PW 4700 (Philips, The Netherlands) or a luminometer LB 9501-1 Lumat (Berthold, Germany) General Methods
Purification of Plasmids for Transfection Plasmids are purified by two rounds of cesium chloride gradient centrifugation or, preferentially, superior purity is achieved by column z8 M. Strauss, S. Hering, L. Lfibbe, and B. E. Griffin, Oncogene 5, 1223 (1990). z9 M. Wirth, R. G r a n n e m a n n , and H. H a u s e r , J. Virol. submitted.
54
VECTORS FOR EXPRESSING CLONED GENES
[5]
chromatography on BD-cellulose as follows: A cleared alkaline lysate is neutralized, phenol treated, and precipitated with ethanol as in the standard procedure. The pellet is dissolved in TE buffer [10 mM TrisHCI, 1 mM ethylenediaminetetraacetic acid (EDTA), pH 7.6]; RNA and protein are selectively precipitated by addition of ammonium acetate to a final concentration of 2.3 M. The precipitate is removed by centrifugation at 6000 rpm/min for 10 min at 4°. The DNA is precipitated from the supernatant with 70% (v/v) ethanol. After centrifugation the pellet is dissolved in 3 ml TE buffer with 0.3 M NaCI (pH 8). Two milliliters of BD-cellulose in TE buffer plus 0.3 M NaCI is filled into a 4- to 5-ml plastic column or syringe and the DNA solution is applied. The column can be spun; however, for maximal yields it should be run by gravity and the flow-through should be applied a second time. After washing with five column volumes of the same buffer the plasmid DNA is eluted with TE buffer plus 1 M NaCI and precipitated with 70% (v/v) ethanol.
Transfection Procedures For transfection of fibroblasts we generally use a modification of the standard calcium phosphate coprecipitation method. 23'3° The highest efficiencies (approximately 10 -3) a r e obtained as follows: 10 /zg of plasmid DNA in 220/xl of TE buffer plus 30/zl of 2 M CaCI2 are mixed in a test tube, and 250 t-d of 2 × HBS [50 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), 280 mM NaCI, 1.5 mM sodium phosphate (equal amounts of mono- and dibasic), pH 6.96] is added dropwise with simultaneous vortexing. The precipitate is directly added to 5 ml of culture medium containing 10% (v/v) fetal calf serum in 25cm 2 tissue culture flasks with cells that have been seeded the day before. For transfection of myeloma cells 31 8/xg of plasmid DNA is precipitated with ethanol, resuspended in 250 ~1 of RPMI 1640-25 mM HEPES (pH 7.15), 250 /zl of RPMI 1640/HEPES containing 1 mg/ml DEAEdextran is added to the DNA with vortexing, and the mixture is incubated with the cells on 10-cm petri dishes for 30 min. After washing with RPMI 1640, 7 ml of medium with serum plus 0.1 mM chloroquine are added for 3.5 hr. Cells are washed again and supplied with fresh medium. Electroporation of 10 7 cells/ml in culture medium with 10% (v/v) fetal calf serum is generally performed on ice with 10-50/xg of linear DNA using the following conditions for different cell types: 1.5 kV/cm 30 F, L. Graham and A. van der Eb, Virology 52, 456 (1973). sl A. Lieber, M. Teppke, G. Herrmann, and M. Strauss, FEBS Lett. 282, 225 (1991).
[5l
HIGH-LEVEL GENE EXPRESSION BY T7 R N A POLYMERASE
55
for 3.5 msec with Sp2/0 and 3.0 kV/cm for 3.5 msec with CHO or Ltkcells.
Reporter Gene Assays Chloramphenicol Acetyltransferase Activity. Our assay is a modification of the original procedure of Gorman et al. 32 as described previously .23 Luciferase Activity. Cells are trypsinized and incubated with medium plus 10% (v/v) calf serum for 30 min and are subsequently washed with phosphate-buffered saline (PBS). Then 106 cells are lysed in 1 ml buffer containing 25 mM Tris-phosphate (pH 7.8)/8 mM MgC12/1 mM dithiothreitol (DTT)/I% (v/v) Triton X-100/I% (v/v) bovine serum albumin/15% (v/v) glycerol/0.4 mM phenylmethylsulfonyl fluoride (PMSF). Cell debris is removed by centrifugation. The enzyme reaction mixture contains 50/xl lysate, 2.5 ~1 of 10 mM ATP, and 5 p~l of 1 mM luciferin. The latter is added 10 sec before introducing the reaction tube into the scintillation counter. Photon emission is counted for 10 sec. Photon emission is counted immediately with the luminometer. Between 10 sec and 5 min after addition of luciferin the photon emission does not decrease significantly. There is a linear dependence of emission from the enzyme concentration between 102 and 10 7 counts. Growth Hormone Detection. An enzyme-linked immunosorbent assay (ELISA) is used in which a monoclonal anti-hGH antibody (1:2000 in PBS) is fixed to a microtiter plate by overnight incubation at 4°. After washing three times with water and with PBS plus 0.05% (v/v) Tween 20 the cavities are blocked with 150 /~i of 0.5% (v/v) bovine serum albumin in PBS at 37° for 1 hr. Then 50 /~1 of cell culture supernatant is added to the coated cavities, incubated at 37° for I hr, and washed off with PBS plus 0.05% (v/v) Tween 20. After adding 50 /~I of a rabbit anti-hGH antiserum (1 : 5000), incubating at 37° for 1 hr, and washing, 50/~1 of peroxidase-conjugated goat anti-rabbit antibodies is added. Following incubation at 37° for 1 hr and extensive washing with water, 100 p,l of the staining solution is added containing 25 mg/ml o-phenyldiamine (OPD) in 0.1 M phosphate-citrate buffer (pH 5.0). The color intensity is measured in a photometer. A commercial hGH (Serono, Switzerland) is used as a standard. Linear response is observed between 0.5 and 10 ng/ml. Detection of Tissue Plasminogen Activator. The tissue plasminogen activator (tPA) protein was determined using an ELISA exactly as described for hGH. 32 C. Gorman, L. F. Moffat, and B. Howard, Mol. Cell. Biol. 2, 1044 (1982).
56
VECTORS FOR EXPRESSING CLONED GENES
[5]
Experimental Procedures Cloning and Isolation of Mutant 77 Promoters Mutant Oligonucleotide Synthesis. The sequence of the T7 promoter with additional flanking restriction sites is synthesized using a modification of the mutagenic procedure published previously. 31 To 2.9 ml of the main phosphoamidite solution (0.13 M) in each position 78/xl of each of the other three amidite solutions is added, resulting in an impurity of 8% for each solution. The oligonucleotide mixture is vacuum dried and dissolved in TE buffer. One microgram of the oligonucleotide mixture is annealed to 1/xg of a 18-mer primer completely complementary to the 3' end of the promoter/linker sequence in 10/.tl of TE by heating to 95 ° and slow cooling to room temperature. The tube is briefly centrifuged. Double-strand synthesis is performed in a total volume of 25/zl with I0 U of Klenow fragment of DNA polymerase in the presence of a 0.4 mM concentration of all four deoxynucleotide triphosphates and 1 /zCi of [32p]ATP for 1 hr at room temperature. Double strands are purified by gel electrophoresis in a 6% (w/v) polyacrylamide gel, excission of the top band, maceration, and elution in 0.5 M ammonium acetate at 37° overnight. The eluate is applied to a C18 column (SepPak; Waters Chromatography Division, Millipore, Milford, MA), the latter is washed with water, and elution is done with I vol of 60% (v/v) methanol. The eluate is dried in a vacuum centrifuge (SpeedVac) and redissolved in TE buffer. HeLa Nuclear Extracts. HeLa cells are grown in roller bottles to a total of 101° cells. Extracts are prepared according to the procedure described by Manley, 33 which is not outlined here. Following precipitation with (NH4)zSO 2 the extract is dialyzed against a 1000-fold volume of 20 mM Tris-HCl (pH 7.9)-20% (v/v) glycerol-0.2 mM EDTA-10 mM 2-mercaptoethanol-0. I M KC1-0.5 mM PMSF. Extracts with a protein concentration of 15 mg/ml are stored in aliquots in liquid nitrogen. Promoter Adsorption to Nuclear Proteins. Salt conditions must be optimized. From our experience we recommend the following composition: Tris-HCl (pH 7.9), 12 mM KCI, 50 mM MgClz, 10 mM 2-Mercaptoethanol, 10 mM EDTA, 0.2 mM Glycerol (15%, v/v) 33 j. L. Manley, in "Transcription and Translation: A Practical Approach" (B. D. Hames and S. J. Higgins, eds.), p. 91. 1RL Press, Oxford, England, 1987.
[5]
HIGH-LEVEL GENE EXPRESSION BY T7 R N A POLYMERASE
57
The incubation mixture contains this buffer and the following additives in a total of 20 txl: H e L a extract (10-20/xg) diluted in dialysis buffer Escherichia coli DNA (2/xg) and 1 /xg hDNA After preincubation at room temperature for 5 min, 20 ng of end-labeled double-stranded mutant oligonucleotides is added. Following incubation at room temperature for 20 min and addition of 2/xl electrophoresis sample buffer the sample is run in a 5% (w/v) nondenaturing polyacrylamide gel at low ionic strength (6.8 mM Tris-HC1, pH 7.9/1 mM EDTA/3.3 mM sodium acetate) with buffer circulation. The wet gel is exposed to X-ray film for 30 min and the band corresponding to unbound oligonucleotides is excised. After elution with 0.5 M ammonium acetate (1.5 ml) at 37° overnight and purification over a C18 column, 50/zg of proteinase K and 1 mM CaC12 are added. The treatment is stopped after 15 min at 37° by addition of 5 mM EDTA and an equal volume of phenol. The oligonucleotides are precipitated with 10 vol of ethanol in the presence of 5 /xg dextran T-500. After amplification by polymerase chain reaction (PCR) the incubation with HeLa extract is repeated three times. Amplification by Polymerase Chain Reaction. The oligonucleotides are redissolved in 30/xl H20 and 300/zg total of both primers corresponding to the nonmutated flanking sequences of the T7 promoter is added together with 12.5 ~1 of a 20% (v/v) Chelex 10034solution. After heating to 95 ° for 5 min and cooling on ice, Taq buffer, 2 mM concentrations of all four deoxynucleotides, and I U Taq polymerase are added. Thirty cycles (1 min at 94 °, 45 sec at 50 °, 30 sec at 74 °) are performed with a 10-rain postincubation at 74 °. After estimation of the amount of oligonucleotides, 20 ng is treated again with HeLa extracts as above. Binding to T7 RNA Polymerase. Conditions for incubation of oligonucleotides with T7 RNA polymerase are exactly as described above for HeLa extracts, using 250 U of enzyme instead of extract. Following electrophoresis the retarded band was excised, purified, and amplified as described. Cloning of Mutant Promoters. One microgram of the polymerasebinding mutant oligonucleotides is cleaved with 100 U of both restriction endonucleases having sites in the flanking sequences (EcoRI and XbaI in our case) and ligated to the human growth hormone gene in pUC19. Following transformation bacterial colonies are screened for inserts by using end-labeled PCR primers. Transcription in Vitro. Plasmid minipreparations can be used after 34 j. Singer-Sam, R. L. Tanguay, and A. D. Riggs, Amplifications 3, 11 (1989).
58
VECTORS FOR EXPRESSING CLONED GENES
[5]
extensive RNase treatment. Individual mutant promoter plasmids are tested for transcription with T7 polymerase in the presence or absence of HeLa cell extract by incubating in the following reaction mixture (25/A): HeLa extract (8 mg/ml) or dialysis buffer, 12.5/xl MgC1z, 10 mM RNasin, 2 U ATP, GTP, and UTP, 500/zM [32p]CTP (0.5/zCi), 50/xM T7 RNA polymerase, 50 U Plasmid DNA, 1/~g After incubation at 30 ° for I hr, samples are applied to Whatman (Clifton, NJ) GF/C filters, precipitated with 10% trichloroacetic acid (TCA), washed with TCA, ethanol, and acetone, and measured in a liquid scintillation counter. Cellular Expression Assay. Both CHO and CHO185 cells are transfected with a mixture of 20/zg test plasmid, 2/xg pRSVLuc, 35 and 1 /~g pT7neo using the protocol described above. Assays for hGH secretion and luciferase (transfection standard) are performed after 3 days and one-tenth of the cells are subjected to selection with 600/zg/ml G418 (Geneticin) to generate pools of about 100 colonies for testing stable expression levels. We have analyzed 61 mutant promoters) 6 The expression levels obtained for some of them are given in Table I. The higher transcription of the mutants in vitro in the presence of HeLa extracts reflects the reduced binding of inhibitory factors whereas the reduced expression in CHO cells in vivo is probably due to loss of polymerase II binding. Selection of Cell Clones Carrying T7 R N A Polymerase Several selectable marker genes were cloned downstream of the wildtype or mutant T7 promoter and have been tested for selectability in both CHO and CHO185 cells as well as after cotransfer with pMTT7N. From our experience we recommend using the neo gene. It is actually expressed from the wild-type T7 promoter only in cells expressing T7 polymerase. In contrast, other selectable marker genes can be expressed at a functional level by cellular RNA polymerases from the wild-type T7 promoter. Data of two experiments using the neo gene and selection with Geneticin are given in Table II. In cases in which the rapid establishment of the T7 expression system 35 j. R. de Wet, K. V. Wood, M. DeLuca, D. R. Helinski, and S. Subramani, Mol. Cell. Biol. 7~ 725 (1987). 36 A. Lieber, V. Sandig, and M. Strauss, Eur. J. Biochem. submitted (1992).
[5]
H I G H - L E V E L G E N E EXPRESSION BY TABLE
T7
RNA
59
POLYMERASE
I
PROPERTIES OF MUTANT T 7 PROMOTERS
Promoter
Sequence -18
Wild type Mutant 13 Mutant 44 Mutant 57 Mutant 68 Mutant 86
-10
+1
Transcription"
Expression ~'
in vitro (%)
in vioo I%)
-HeLa
+HeLa
CHO185
CHO
100 70 82 73
10 46 40 46
115 64 70 54
100 6 20 4
50 80
43 55
37 67
8 4
+8
TTAATACGACTCACTATAGGGAGATA ........ G ............ G .... -C - C ................... T- ..... T ............ C - T ..... ...... T ..... C ........ TT- - ..... T ............ C .......
" Transcription by T7 polymerase in vitro was done in the absence or presence of H e L a nuclear extracts. Counts obtained for the wild-type promoter in the absence of extract were taken as 100c;~. h Stable expression of human growth hormone was determined in populations of 100-200 colonies obtained after cotransfer with pT7neo and selection with Geneticin. The level obtained for the wildtype promoter in CHO cells is the 100% value.
in a particular cell type is required we suggest a cotransfer of the gene of interest under control of the mutant T7 promoter together with pMTT7N and pT7neo at a ratio of 20 : ! : 1. In this case 100% of the resulting colonies express T7 polymerase and more than 90% will express the gene of interest. Physical linkage of the gene of interest to pT7neo would guarantee
T A B L E II T 7 R N A POLYMERASE-DEPENDENT SELECTION FOR neo GENE EXPRESSION Colonies a Plasmid pUCII8
pT7neo pT7neo + pMTT7N
C H O cells
C H O 1 8 5 cells
0/0 2/0 100/60
0/1 200/300 ND
T w o d a y s a f t e r t r a n s f e c t i o n w i t h 6 / . t g o f the respective plasmid, the calcium phosphate techn i q u e w a s u s e d to s e e d 1 × 105 cells into 10-cm d i s h e s w i t h D M E M plus 5% ( v / v ) fetal c a l f s e r u m plus 600 p . g / m l G e n e t i c i n . S e l e c t i o n w a s f o r 2 weeks, with medium changes and colony counts m a d e e v e r y third day.
60
VECTORS FOR EXPRESSING CLONED GENES
[5]
expression in 100% of the colonies. This can also be achieved by using bicistronic structures (see below). However, we observed a significantly lower colony number in this kind of cotransfection experiment compared to experiments with cells already expressing T7 polymerase. In most clones, the level of T7 polymerase expression is considerably lower than that expected for metallothionein promoter constructs. It appears that there is a counterselection against cells expressing high levels of T7 polymerase. This conclusion is also supported by the observed difficulty in establishing stable clones with pCMVT7N (Fig. 1), in which T7 RNA polymerase expression is driven by the strong CMV promoter. If high-level expression is the final goal we recommend the use of a two-step protocol: pMTT7N is cotransfected with pT7neo at a ratio of 10 : 1 and cells are selected with the desired concentration of Geneticin depending on the cell type, usually between 400 and 800 ~g/ml. The pool of colonies can be used directly for the second transfection with the respective gene under control of a mutant T7 promoter together with either pT7mpac, pT7mhyg (see Vector and Oligonucleotides, above) or any other marker gene. Bicistronic constructions can also be useful in this case (see below).
Alternative Strategies for T7-Driven Gene Expression If the expression of a particular gene product is disadvantageous for the host cell or should be switched on at a fixed time point it is desirable to have a cell line carrying a silent gene. Two different ways are imaginable for activation of a gene under control of a mutant T7 promoter. First, a tightly regulated promoter could be used. We are currently exploring the usefulness of bacterial operator/repressor systems for this purpose. The use of eukaryotic regulatory elements such as hormone-response elements turned out to be inefficient in this context. The second alternative is the use of retroviral vectors to introduce the T7 polymerase gene efficiently into nearly 100% of the cells. To this end we have developed vectors based on pM6pac expressing the T7 polymerase gene from the retroviral long terminal repeat (LTR) or from CMV or SV40 promoters (Fig. 1). The experimental strategy is as follows. A vector containing the particular gene downstream of the T7m promoter is cotransfected with a selectable marker. After selection individual clones are analyzed for the presence of integrated and intact gene constructions by Southern blotting or PCR. The cells are then infected with a retroviral vector carrying the polymerase gene. We have tested this protocol for expression of the neo gene. Linear pT7neo (6 tzg) was transfected by the calcium phosphate technique into
[5]
HIGH-LEVEL GENE EXPRESSION BY T7 R N A POLYMERASE
61
5 x 105 NIH 3T3 cells together with 0.5 p~g of pY337 and selection with 50 /zg/ml of hygromycin was started the day after. Resulting colonies were pooled and tested for growth in 400/zg/ml of Geneticin. No survival of cells was observed. The cells were then infected with the retroviral vector pM6SVT7N (Fig. 1), which has been packaged by transfection of GP + E86 cells and was selected with 4/zg/ml of puromycin. The titer of the virus has been estimated on N I H 3T3 cells to be 1 x 105. The cells carrying the silent neo gene (105/5-cm dish) were infected with different aliquots of M6SVT7N virus stock (0.01, 0.1, and 1 ml). One day later cells were seeded at a concentration of 1000/5-cm dish in triplicate and selected with 400 /~g/ml of Geneticin. Parallel dishes without selection allowed the determination of plating efficiency. With the highest amount of virus (1 ml) the number of colonies in Geneticin equaled the plating efficiency without selection, suggesting a nearly 100% induction of expression of the neo gene. 38
Applications High-Level Foreign Gene Expression Stable Expression of the Genomic Human Growth Hormone Gene. All previous experiments for high-level expression have been performed with the wild-type T7 promoter. Because we have isolated optimized mutant promoters only recently, we have at present no long-term expression data. However, the preliminary data suggest that the levels of expression from the wild-type and mutant promoters are comparable. Here we describe our protocol, which leads to high-level expression of human growth hormone. A plasmid pGGH2.2 is constructed by cloning a 2.1-kbp genomic human growth hormone gene between the BamHI and EcoRI sites of pGEM2. Ten micrograms of linearized plasmid is cotransfected with 0.5 tzg pSV2neo 39 into 3 x 105 polymorphonuclear (PMN) mouse cells (clone A5) by the calcium phosphate technique as described above. After an overnight incubation with the DNA precipitate, cells are washed and supplemented with fresh medium. After 2 days cells are trypsinized, divided 1 : 6 into new dishes, and subjected to selection with 400/zg/ml of Geneticin. Medium is changed every third day for a period of 3 weeks. Colonies are trypsinized all together and seeded into 96-well microtiter plates at a dilution of 1 cell per every second well. Medium with Geneticin 37 K. Blochlinger and H. Diggelmann, Mol. Cell. Biol. 4, 2929 (1984). 38 W. Sommer and M. Strauss, manuscript in preparation (1992). 39 p. j. Southern and P. Berg, J. Mol. Appl. Genet. 1, 327 (1982).
62
VECTORS FOR EXPRESSING CLONED GENES
[5] kD
-
97
-
66
-
45
- 24 -4----
hGH
FIG. 3. Electrophoretic pattern of immunoprecipitated hGH produced by a PMN cell clone harboring a T7hGH plasmid. Lanes 1 and 2 are immunoprecipitates from extracts of 10 6 cells, lanes 3 to 8 are precipitates from 100/.d of culture supernatant. Lane 1 is fresh extract, lane 2 is extract incubated at 37° for I hr, lanes 3 to 5 are successive 24-hr supernatants of logarithmic cells, and lanes 6 to 8 represent the same culture supernatant of a confluent monolayer after 24, 48, and 72 hr. Immunoprecipitation was done by subsequent incubation with antiserum and protein A-Sepharose. The precipitates were run in a 12% (w/v) polyacrylamide gel.
is changed once a week. After 3 weeks supernatants are assayed for growth hormone secretion by the ELISA described above. High producers (more than 10/zg/ml hGH) are recloned to guarantee later stability of the clones. Using this protocol we succeeded in isolating several clones producing considerably more than 10/xg/ml/106 cells/day from approximately 250 primary colonies, with 2 clones producing 20 and 30 tzg/ml. These levels are slightly above those obtained using CMV or metallothionein promoter constructs. 36 However, the percentage of good producers is much lower (about 1 out of 400) in the latter case. The yield of growth hormone is best if the medium is changed every day and better in the exponential growth phase compared to a stationary culture. A typical pattern of growth hormone in cell extracts and in the medium after immunoprecipitation is shown in Fig. 3.
[5]
63
HIGH-LEVEL GENE EXPRESSION BY T7 R N A POLYMERASE AAYA~
AA'TAAA
A VA
#"~
~ T ~
~TJO~
B PA
P'A
C
iI
T7prom.
V. rnouw
CI~ huml~'l
VKm~.
,
,
~-,khurrmn
FIG. 4. Maps of bi- and polycistronic constructions driven by a T7 promoter,
Stable Expression of Tissue Plasminogen Activator cDNA. The cDNA for human tissue plasminogen activator is cloned into pGEM downstream to the T7 promoter. The linear plasmid (6 tzg) is cotransfected with pSV2neo (0.5 /~g) into CHO185 cells. After 3 days, selection with 600 /xg/ml of Geneticin is started. Three weeks later colonies are pooled and cloned in microtiter plates as described for growth hormone expression. Clones producing more than 2 /~g/10 6 cells/day are recloned three times before stable expression levels can be detected. The highest level obtained in our experiments was 5 /zg/10 6 cells/day. Expression of Bi- and Polycistronic mRNA Selection of Expressing Clones. The coding sequence of the neo gene is cloned downstream to the gene of interest. The presence of a polyadenylation signal does not interfere with transcription of the bicistronic mRNA. We used fusion with the tPA cDNA as a model. A plasmid pGtPAAneo is constructed (Fig. 4A) and transfected into PMN cells (clone A5). Selection is performed using 400/zg/ml of Geneticin. The number of colonies obtained is about 10-fold lower as compared to transfections with pSV2neo (20 colonies/5-cm dish). However, all colonies give rise to stable clones expressing moderate to high levels of tPA. No colonies can be obtained in CHO cells due to the function of the polyadenylation signal as a transcriptional terminator. Alternatively, the luciferase coding sequence can be fused downstream to the gene of interest to allow for rapid quantitation of gene expression. We use a tPA-luc fusion gene. The plasmid pGtPAAL (Fig. 4B) is cotrans-
64
VECTORS FOR EXPRESSING CLONED GENES
[5]
fected with pT7mpaC into CHO cells expressing T7 polymerase. Selection is carried out with 4/.tg/ml ofpuromycin. Colonies are tested for luciferase expression, which can be detected at the level of 1-10 cells. However, the luciferase activity is decreased by a factor of five when the cistron is expressed in the second position instead of in the first position. The level of tPA synthesis correlates very well with the luciferase activity. No luciferase activity can be detected in CHO cells with this construction. As soon as a semiquantitative in situ assay for luciferase activity is available this kind of gene fusion will allow rapid identification of producer clones. Expression of Chimeric lmmunoglobulins. A plasmid pT7Ig may be constructed (Fig. 4C) in which the genomic sequences for variable regions of a mouse antibody to pig transferrin (a gift of Dr. S. Deev, Moscow) are fused to the genomic sequences for the constant regions of human IgE. The plasmid, containing two internal polyadenylation signals (which do not influence transcription by T7 polymerase), is cotransfected by electroporation with pT7mpac into Sp2/0 myeloma cells previously selected for expression of T7 RNA polymerase. After selection and subcloning in 96well microtiter plates secretion of IgE is determined using a commercial ELISA. Average levels of 100 ng/ml of IgE are detectable. Extracts are prepared and subjected to electrophoresis, and the gel is blotted to nitrocellulose and probed with anti-IgE. A 75-kDa heavy chain and a 25-kDa light chain are found. Using a solid-phase adsorption assay for pig transferrin, specific binding activity of the chimeric antibody is confirmed. Northern blot analysis shows the presence of a major transcript of about 8 kb. 4°
Synthesis of Antisense RNA and Ribozymes Recent applications of the T7 expression system in o u r laboratory have been directed toward the high-level synthesis of antisense RNA and ribozymes. The principle is described in brief: The T7 system allows expression of several thousand up to 30,000 RNA molecules per cell. For mRNAs of low or moderate abundance this level of antisense RNA might be sufficient to knock out the mRNA. We have tested this assumption by expression of a 365-bp fragment from the first exon of the human Rb-1 gene. We generated stable cell lines from CHO cells and primary human fibroblasts. In both systems no synthesis of the RB protein was detectable. Whereas the first were converted into tumorigenic cell lines, the latter have a dramatically shortened cell cycle.41 The tumorigenic cell lines might be immortalized, which is currently under investigation. 4o A. Lieber and M. Strauss, manuscript in preparation. 41 M. Strauss, S. Hering, A. Lieber, G. Herrmann, B. E. Griffin, and W. Arnold, Oncogene, in press (1992).
[5]
HIGH-LEVEL GENE EXPRESSION BY T7 R N A POLYMERASE
65
,p --- UAeC C eeUCUC CA UCUAU---"Rm ~ n GGAGGACAUGCGGGGC~ ACUGUCGUUGCAGAUA
AA %GA AGcGAGU ~'
GC AGua
FIG. 5. Structure of a ribozyme construction directed to the mRNA of hGH.
For mRNAs of higher abundance the use of hammerhead ribozymes 42'43 might be more efficient than using simple antisense RNA. We have designed a ribozyme construction against the mRNA for hGH (Fig. 5). The hammerhead structure and the flanking complementary sequences were synthesized as three overlapping oligonucleotides, annealed, filled in with Klenow polymerase, and cloned into the EcoRI site of pGEM1 together with the coding sequence of the neo gene. The resulting plasmid, pRZGHneo, was used to transfect PMN cells already expressing hGH from the T7 promoter. Following selection with Geneticin clones were analyzed for hGH secretion. The levels varied between 25 and 2.5% of the original hGH synthesis, with the majority of clones producing very low levels of hGH. Discussion and Comments The heterologous gene expression system described here has several advantages over others. However, several problems must be considered. First, the T7 promoter is not exclusively transcribed by T7 polymerase within the nucleus of mammalian cells. Second, the levels of expression obtained with the T7 system are not significantly higher than those obtained with strong eukaryotic promoters under optimal conditions. An third, high-level expression of the T7 RNA polymerase is disadvantageous to the host cell. The considerable expression level caused by the cellular RNA polymerase with the wild-type promoter does not significantly influence the expression by T7 polymerase. Thus, we may recommend the use of the wildtype promoter in most cases where high-level stable gene expression is 42 A. C. Forster and R. H. Symons, Cell 49, 211 (1987). 43 C. C. Sheldon, A. C. Jeffries, C. Davies, and R. H. Symons, in "Nucleic Acids and Molecular Biology" (F. Eckstein and D. M. J. LiUey, eds.), Vol. 4, p. 227. SpringerVerlag, Berlin, 1990.
66
VECTORS FOR EXPRESSING CLONED GENES
[6]
required. The advantage over other stable expression systems is the ease of selecting producer clones, not the height of expression. The expression levels are always lower in stable systems compared with the lytic viral ones. We tried to make use of the excess of T7 polymerase present in the nucleus of some clones by using episomal origin vectors. However, up to now the levels of expression could be enhanced only two- to fourfold. 44 We are currently working on the establishment of regulated systems that might allow a higher level of expression for a short period of time. The chief advantage of the modified T7 system based on mutant promoters is the strong dependence on transcription by T7 RNA polymerase. Thus, the system can be used for all purposes of selective gene expression. Cell lines or even transgenic mice carrying silent genes under the control of a mutant promoter can be established and their expression can be stimulated via subsequent introduction of T7 polymerase, for example, by retroviral vectors. Alternatively, inducible repressor systems could be used to repress either the T7 promoter or T7 polymerase expression. Additional useful applications are the expression of bi- or polycistronic mRNA as well as the efficient expression of antisense RNA or ribozymes. 44 V, Sandig and M. Strauss, unpublished observations 0991),
[6] E x p r e s s i o n V e c t o r s f o r H i g h - L e v e l G e n e E x p r e s s i o n in Dicotyledonous and Monocotyledonous Plants By REINHARD TOPFER, CHRISTOPH MAAS, CHRISTA H(SRICKE-GRANDPIERRE, J E F F SCHELL,
and
H A N S - H E N N I N G STEINBISS
Introduction High-level expression of selectable marker genes as well as that of agronomically important genes is a crucial aspect of plant molecular biology. The first chimeric genes for plant transformation experiments consisted of bacterial antibiotic resistance genes controlled by promoters derived from genes carried by the T-DNA region ofAgrobacterium tumefaciens Ti plasmids. ~-3 A second generation of constructs made use of the L. Herrera-Estrella, M. De Block, E. Messens, J.-P. Hernalsteens, M. Van Montagu, and J. Schell, EMBO J. 2, 987 (1983). M. Bevan, R. B. Flavell, and M.-D. Chilton, Nature (London) 304, 184 (1983).
METHODS IN ENZYMOLOGY,VOL.217
Copyright© 1993by AcademicPress, Inc. All rightsof reproductionin any formreserved.
[6]
GENE EXPRESSION IN DICOTS AND MONOCOTS
67
s t r o n g e r 355 R N A p r o m o t e r o f the cauliflower m o s a i c virus ( C a M V ) , w h i c h in fact has p r o v e d to be o n e o f the m o s t effective and best c h a r a c t e r ized plant p r o m o t e r s . 4-7 This p r o m o t e r has b e e n u s e d to create e x p r e s s i o n v e c t o r c a s s e t t e s s u c h as pDH51,8 p M O N 3 1 6 , 9 and pRT100,~° with w h i c h high levels o f gene e x p r e s s i o n c a n be o b t a i n e d in a variety o f plant tissues. Interestingly, the transcriptional activity o f this p r o m o t e r could be stimulated up to 10-fold b y duplication o f the C a M V 355 e n h a n c e r s e q u e n c e s ~ or b y insertion o f a 42-bp S h r u n k e n 1 e x o n 1 s e q u e n c e j u s t d o w n s t r e a m o f the t r a n s c r i p t i o n start site. 12 M o r e o v e r , it has b e e n s h o w n that gene e x p r e s s i o n c a n also be m o d u l a t e d significantly at the posttranscriptional level. I n c l u s i o n o f introns in the t r a n s c r i p t i o n unit s e e m to influence stability o f the transcript, leading to an e n h a n c e m e n t o f gene e x p r e s s i o n of up to 150-fold. t2-14 T h e use o f different 3' p o l y ( A ) cassettes resulted in differences in gene e x p r e s s i o n o f up to 60-fold. ~5 T h e translational effic i e n c y o f c h i m e r i c transcripts w a s s h o w n to be increased by insertion o f viral u n t r a n s l a t e d leader s e q u e n c e s in the 5' u n t r a n s l a t e d region o f the t r a n s c r i p t i o n unit, resulting in an e n h a n c e d level o f gene e x p r e s s i o n o f up to 3 5 - f o l d ) 6A7 This c h a p t e r d e s c r i b e s sets o f v e c t o r s that are derivatives of the expression v e c t o r c a s s e t t e pRT100, w h i c h uses the C a M V 355 R N A p r o m o t e r in c o m b i n a t i o n with v a r i o u s r e p o r t e r and selectable m a r k e r genes. F u r t h e r m o r e , w e i n t r o d u c e a set o f e x p r e s s i o n v e c t o r s f o r e n h a n c e d gene expression in m o n o c o t y l e d o n o u s plants, especially a g r o n o m i c a l l y i m p o r t a n t ce-
3 R. T. Fraley, S. G. Rogers, R. B. Horsch, P. R. Sanders, J. S. Flick, S. P. Adams, M. L. Bittner, L. A. Brand, C. L. Fink, J. S. Fry, G. R. Galuppi, S. B. Goldberg, N. L. Hoffmann, and S. C. Woo, Proc. Natl. Acad. Sci. U,S.A. 80, 4803 (1983). 4 j. T. Odell, F. Nagy, and N.-H. Chua, Nature (London) 313, 810 (1985). D. W. Ow, J. D. Jacobs, and S. H. Howell, Proc. Natl. Acad. Sci. U.S.A. 84, 4870 (1987). 6 R.-X. Fang, F, Nagy, S. Sivasubramaniam, and N.-H. Chua, Plant Cell 1, 141 (1989). 7 p. N. Benfey and N.-H. Chua, Science 250, 959 (1990). 8 M. Pietrzak, R. D. Shillito, T. Hohn, and I. Potrykus, Nucleic Acids Res. 14, 5857 (1986). 9 S. G. Rogers, H. J. Klee, R. B. Horsch, and R. T. Fraley, this series, Vol. 153, p. 253. I0 R. TOpfer, V. Matzeit, B. Gronenborn, J. Schell, and H.-H. Steinbiss, Nucleic Acids Res. 14, 5890 (1987). 11R. Kay, A. Cahan, M. Daly, and J. McPherson, Science 236, 1299 (1987). ~2C. Maas, J. Laufs, S. Grant, C. Korfhage, and W. Werr, Plant Mol. Biol. 16, 199 (1991). ~3j. Callis, M. Fromm, and V. Walbot, Genes Dev. 1, 1183 (1987). t4 V. Vasil, M. Clancy, R. J. Ferl, 1. K. Vasil, and C. Hannah, Plant Physiol. 91, 1575 (1989). ~51. L. W. Ingelbrecht, L. M. F. Herman, R. A. Dekeyser, M. C. Van Montagu, and A. G. Depicker, Plant Cell 1, 671 (1989). 16S. A. Jobling and L. Gehrke, Nature (London) 325, 622 (1987). 17D. R. Gallie, D. E. Sleat, J. W. Watts, P. C. Turner, and T. M. A. Wilson, Nucleic Acids Res. 15, 3257 (1987).
68
VECTORS FOR EXPRESSING CLONED GENES
[6]
reals. These vectors have been constructed as basic tools applicable for transient gene expression as well as for stable integration of foreign genes into plant genomes. Improved Vector Cassettes for Gene Expression in Dicotyledonous Plants
Expression Vector Cassettes pRTIO0 to pRTI08 General Properties of Vectors. The vectors pRTI00-pRT108 (Fig. 1) are plant expression vector cassettes that permit expression of any gene of interest under the control of the 35S RNA promoter of CaMV and its corresponding polyadenylation [poly(A)] signal. The common feature of this set of vectors is the symmetrical arrangement of restriction sites (HindlII, SphI, and PstI) bordering both regulatory elements, providing the possibility for the simple transfer of a given chimeric construct to other vectors. The various vectors carry different polylinker sequences separating promoter and poly(A) signal sequences, as indicated schematically in Fig. 1 and at the nucleotide level in Fig. 2. With respect to application, some vectors are suitable for translational and others for transcriptional fusions. Vectors Suitable for Translational Fusions. pRT100, pRT102, pRT103, pRT104, pRTI07, and pRT108 carry an NcoI site providing an ATG codon, which can be used as the ATG codon for translational fusions. In pRT103, pRT104, pRTI07, and pRT108 the ATG is embedded in the consensus sequence for optimal ribosome initiation. ~8 Vectors Suitable for Transcriptional Fusions. pRT101, pRTI05, and pRT106 were designed to create transcriptional fusions, pRT105 and pRTI06 carry the pBluescript ~9multiple cloning site in both orientations. Construction ofpRTlO0 to pRTI08. All cloning was performed according to standard protocols as described by Sambrook et al. 2° The intergenic region of the CaMV genome carrying the 35S RNA promoter and the corresponding poly(A) signal were obtained as an 800-bp EcoRI fragment of CaMV Cabb B-D subcloned in phage fdl 1-6.z~An HphI site (at position 7437, referring to CaMV isolate CM184122) separating the promoter and 18 M. Kozak, Nucleic Acids Res. 2, 857 (1984). 19 Stratagene, 11099 North Torrey Pines Road, La Jolla, CA 92037. 2o j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NewYork, 1989. 21 V. Matzeit, Thesis. University of Cologne, Cologne, Germany, 1982. 22 R. C. Gardner, A. J. Howarth, P. Hahn, M. Brown-Luedi, R. Shepherd, and J. Messing, Nucleic Acids Res. 9, 2871 (1981).
:= ~ ,,4
v~ x
z,,,
z
m
x
'0
~, ~ - < [.-,
Y_ ~ W -"-/ ~ /
,--
-~ ~..T~
~ ~., 0
~ ~
> .=
0 "~ "~
6.~
~o
~ =
~,,
,8
X
0 8
0
~
~ < ~
69
'~
70 pRTIO0
VECTORS FOR EXPRESSING CLONED GENES ACctcgag...ggcccatgggcgagctcggtaccc
..........
[6]
ggggatcctctagagtcCG
pRTI01ACctcgag...aattc ...... gagctcggtaccc .......... ggggatcctctagagtcCG pRTI02 ACctcgag...ggcccatgggcgagctcggtaccc .......... gg ............... CG pRT103 ACctcgagtggccaccatgggcgagctcggtaccc .......... ggggatcctctagagtcCG pRTI04 ACctcgagtggccaccatgggcgagctcggtacccccgaattcggggggatcctctagagtcCG pRT105 ACccgagctccaccgcggtggcggccgctctagaactagtggatcccccgggctgcaggaattc gatatcaagcttatcgataccgtcgacctcgagggggggcccggtacccagtcCG pRT106 ACcgggtaccgggccccccctcgaggtcgacggtatcgataagcttgatatcgaattcctgcag cccgggggatccactagttctagagcggccgccaccgcggtggagctcCG pRTI07 ACctcgagtggccaccatgggcgagctccaccgcgggggcggccgctctaEaactagtggatcc cccgggctgcaggaattcgatatcaagcttatcgataccgtcgacctcgagggggggcccgg tacccgggtactcCG pRTl08 ACctcgagtggccaccatgggcgagctccaccgcEgtggcggccgctctagaactagtggatcc cccgggtactcCG FIG. 2. The nucleotide sequences of the multiple cloning site between the 35S R N A promoter and the corresponding poly(A) signal of each of the vectors pRT100 to pRTI08 are given. Nucleotides indicated in capital letters belong to the promoter (AC) and poly(A) signal sequence (CG), respectively. Lower-case letters describe the enti~ multiple cloning site.
poly(A) signal 4 bp downstream of the transcription initiation site was cut, converted to blunt ends with Klenow polymerase in the absence of nucleotides, and further digested with either HinclI (position 7016 of CM 1841) or RsaI (po sition 7639 of CM 1841 ). The promoter was subcloned into the HinclI and XhoI sites (filled in) of a modified pUC19 vector, which carried XhoI and NcoI restriction sites instead of SmaI and KpnI, respectively. The poly(A) signal was subcloned in the HinclI site of pUC18, a3 Promoter and terminator were combined using ScaI (in the ampicillin resistance gene of pUC) and SstI (in the multiple cloning site of pUC), resulting in the expression vector cassette pRT100 (Fig. 1). Sequence analysis of the construct revealed a 3-bp deletion at the 3' end of the promoter-carrying fragment, which most likely occurred during removal of the sticky ends of the HphI cleavage site, while leaving intact the transcription initiation site (Fig. 2). Thus, the CaMV sequence elements in pRT100 correspond to bp 7016-7434 (promoter) and bp 7436-7639 [poly(A) signal] of the CaMV isolate CM1841. 22 In addition to pRTI00, derivatives were constructed with modified multiple cloning sites between promoter and terminator (Figs. 1 and 2). 23 C. Yanisch-Perron, J. Vieira, and J. Messing, Gene 33, 103 (1985).
[6]
GENE EXPRESSION IN DICOTS AND MONOCOTS
71
pRT101 to pRT104 have been described elsewhere. 24 pRT105 and pRT106 were created by insertion of KpnI and SstI linkers, respectively, into either the XhoI or XbaI site of pRT104 (compare with Fig. 2) and replacement of its polylinker by the pBluescript multiple cloning site. To obtain the vector pRT107, a KpnI linker was cloned in the S 1 nuclease-treated XbaI site of pRT104 following insertion of the pBluescript multiple cloning site into the KpnI and SstI sites, pRT108 is pRT107 with a Sinai deletion.
Plant Expression Vectors Based on these expression vector cassettes a number of fusions with different reporter genes have been constructed (Fig. 3), which were then used to investigate transient gene expression 25-z7or to establish transgenic plants. To use these expression vector cassettes for further constructions, restriction sites for cloning were chosen in such a way that a minimum of sites were maintained between promoter, structural gene, and poly(A) signal sequence. As indicated below, the sticky ends of a number of restriction fragments created during the cloning procedure were converted to blunt-ended fragments by S1 nulcease treatment prior to ligation. 1. The dhfr gene, conferring methotrexate resistance, was excised from plasmid pFR400 z~ using XbaI/NsiI and ligated with the expression vector pRT100, which was restricted with NcoI/XbaI. The sticky ends of these restriction fragments were removed by SI nuclease treatment and cloning resulted in the vector pRTlOOdhfr. 2. The hpt gene, coding for hygromycin resistance, was derived as a BamHI fragment from pHyml329 (B. Nelsen-Salz, unpublished results, 1982, which was treated with S1 nuclease. This fragment was cloned into the vector pRT100 (restricted with NcoI/XbaI) and treated with SI nuclease, leading to the plasmid pRTlOOhpt. 3. The choramphenicol acetyltransferase (cat) gene of pCAP2123o was subcloned as a BamHI fragment into both pRT101 and M13mp9, leading to pRTlOlcat 25 and M13mp9cat, respectively. An Ncol site at the first .,4 R. TOpfer, V. Matzeit, B. Gronenborn, J. Schell, and H.-H. Steinbiss, Nucleic Acids Res. 14, 5890 (1987). 25 M. Pr61s, R. T6pfer, J. ScheIl, and H.-H. Steinbiss, Plant Cell Rep. 7, 221 (1988). 26 R. T6pfer, M. Pr61s, J. Schell, and H.-H. Steinbiss, Plant Cell Rep. 7, 225 (1988). .,7 R. T6pfer, B. Gronenborn, J. Schell, and H.-H. Steinbiss, Plant Cell 1, 133 (1989). 28 C. C. Simonsen and A. D. Levinson, Proc. Natl. Acad. Sci. U.S.A. 80, 2495 (1983). 29 Originally, the HPT coding sequence is derived from pVU 1011 (P. J. M. van den Elzen), which is equivalent to pLG90 [P. J. M. van den Elzen, J. Townsend, K. Y. Lee, and J. R. Bedbrook, Plant Mol. Biol. 5, 299 (1985)] but carries ATA instead of ATG in the untranslated 5' leader. 30 j. Velten and J. Schell, Nucleic Acids Res. 13, 6981 (1985).
72 HSPHII
VECTORS FOR EXPRESSING CLONED GENES
EV X Sc
Ss
---~m:~m~m-~'~ I pRT100dhfr
HSPHII
[6]
PSH
r////.g---
I
EV XSm E P
PSH
F/////2-
pRT'lOOhpt
HSPHII
EVXBN
BN Sc
PSH
pRTlO3cat HR
HSPHII
EVXBN
~VI~
---~.'.~-~JH~ii~iL~
EV
I
PSH
HII
I
~////2-
pRTlO3gus
HSPHII EVXBN P ~mamHwJ~.i( I
SN
II
PSH p-///,Y,L_.
pRTlO3neo HR
HSPHa
EVXBH
pRT103pat
S~ SS K
PSH ~bp
Fro. 3. Scheme describing the various expression vectors based on expression vector cassettes (Fig. 1). pRTI00 and pRTI03 were used to insert the coding sequences for dehydro-
folate reductase (DHFR), hygomycin phosphotransferase (HPT), chloramphenicol acetyltransferase (CAT), /3-glucuronidase (GUS), neomycin phosphotransferase (NPT II), and phosphinotricin acetyltransferase (PAT). In all of these constructs the cloning sites between the coding sequence and poly(A) signal have been deleted. Restriction endonuclease cleavage sites are abbreviated: B, Bali; E, EcoRI; EV, EcoRV; H, HindllI; HII, HinclI; K, KpnI; N, NcoI; P, PstI; S, SphI; Sa, Sail; Sc, ScaI; Sin, Smal; Ss, SstI; X, XhoI. The bar indicates 100 bp.
[6]
GENE EXPRESSION IN DICOTS AND MONOCOTS
73
ATG codon of the CAT-coding region in M 13mp9cat was introduced using site-directed mutagenesis. 3~The modified cat gene was then inserted into pRTI03 using NcoI and BamHI. Finally, the remaining restriction endonuclease cleavage sites for BamHI and XbaI were removed using S1 nuclease, resulting in plasmid pRTlO3cat. 26 4. An NcoI/EcoRI fragment of pRAJ275,3z carrying the/3-glucuronidase (gus) gene, was cloned into pRT104. Deletion of the restriction sites EcoRI, BamHI, and XbaI was achieved by using S I nuclease treatment, thus creating pRTlO3gus. 26 5. pRTlO3neo z6 resulted from cloning of an NcoI (partially digested)/ BamHI fragment of pRT 100neo, carrying the neomycin phosphotransferase (neo) gene, into pRT 103 followed by removal of the remaining BamHI and XbaI site with S1 nuclease. 6. The bialaphos resistance (bar) gene from pGSFR133 was subcloned into pRT104 as an NcoI/BamHI fragment, subsequently digested with BamHI and XbaI, and the 5' protruding ends converted to blunt ends by S1 nuclease treatment, resulting in the plasmid pRTlO3pat. These vectors (except for the cat vectors) have also been used to create a related set of expression vectors: pRT55, pRT66, pRT77, pRT88, and pRT99 (Fig. 4). They are based on a pUC18 derivative whose PstI site has been substituted by NsiI and which carries a chimeric neomycin phosphotransferase II (NPT II) gene inserted into the MaeI site at position 1554 of pUC18. The NPT II gene of this basic construct, pRT99, 34 has been replaced by the structural genes for dehydrofolate reductase (DHFR--pRT55), hygromycin phosphotransferase (HPT--pRT66), phosphinotricin acetyltransferase (PAT--pRT77), and fl-glucuronidase (GUS--pRT88) using XhoI and PstI. These vectors are suitable for expression and transformation studies and they allow the possibility of cloning a fragment of DNA into the multiple cloning site, taking advantage of the blue/white selection via the lacZ system as in pUC plasmids. 23 Improved Vectors for Gene Expression in Monocotyledonous Plants As already described, the CaMV 35S RNA promoter is highly active in dicotyledonous plants. However, use of this promoter in monocotyle31 W. Kramer, V. Drutsa, H.-W. Jansen, B. Kramer, M. Pflugfelder, and H. Fritz, Nucleic Acids Res. 12, 9441 (1984). 32 R. A. Jefferson, Plant Mol. Biol. Rep. 5, 387 (1987). 33 M. De Block, J. Botterman, M. Vandewiele, J. Dockx, C. Thoen, V. Gossel& N. Rao Movva, C. Thompson, M. Van Montagu, and J. Leemans, EMBO J. 6, 2513 (1987). 34 R. T6pfer, J. Schell, and H.-H. Steinbiss, Nucleic Acids Res. 16, 8725 (1988).
~::~ ~.~ ~
0", I'ne
0~0 ~
In
.i..-
x'
0..
-R
I00
,,e
I.-- u'~
l:'--
re"
,,.., ~
.
i
,.o~
4.
I--n,-
.a .,/
,"~w
C 0
"~
=
~, . _
o
r,-
,..,~
. o .~
< ~ ~ ~ ~ .-~ 0
I
[6]
GENE EXPRESSION IN DICOTS AND MONOCOTS
75
donous plants revealed that gene expression from the CaMV 35S RNA promoter is only 0.1-1% of the level that can be achieved in dicotyledonous plants. Therefore, to increase gene expression in monocotyledonous plants, regulatory elements of the maize Shrunken 1 gene 35 were inserted into the expression vector pRTlOlcat 25(Fig. 5), resulting in the cassettes ~2 described in detail in the following section.
Vector Construction The constructs made use of sequences from the 5' untranslated region of the maize Shrunken 1 gene (Sh 1). A HinclI Sh 1 intron 1 fragment ( + 43 to + 1084) was isolated from a Shrunken 1 promoter clone containing sequences from - 1076 to + 1084 (pSP1076 + 108412). This Sh 1 intron 1 was inserted into the untranslated leader of the chimeric gene of pRT 101 cat by using the unique SmaI site, leading to the construct pRT-int/s-cat (Fig. 5). A sequence from the untranslated Sh 1 exon 1 ( + 4 C C C T C C C T C C CTCCTCCATT GGACTGCTTG CTCCCTGTT+42CCC) was synthesized and inserted into the Sinai site of pRTlOlcat, giving rise to the construct pRT-ex/s-cat (Fig. 5A). The 3'-CCC of the inserted oligonucleotide restored the SmaI site in pRT-ex/s-cat, which was used to insert the HinclI Sh 1 intron I fragment (+ 43 to + 1084) isolated from the Shrunken 1 clone pSP1076 + 1084, thus giving rise to the construct pRt-ex/s-int/s-cat (Fig. 5A).
Expression Studies To analyze these constructs protoplasts of a barley (Hordeum vulgare L. cv. 'Golden Promise '36) and tobacco suspension cell line (Nicotiana tabacum L. cv. W 38; H.-H. Steinbiss, unpublished observations) were used. Preparation of protoplasts from barley and tobacco was essentially as described for maize 37 except that an osmolarity of 720 mOsm was used. Transfection and analysis of CAT activity was as described previously.12'38 Transfection was carried out by using 25/zg plasmid DNA, 100/,~g sonicated calf thymus DNA, polyethylene glycol (PEG 1500) [25% (v/v) PEG, 0.1 M MgClz, pH 6.0], and 1 × 106 protoplasts. Expression was assessed after 40 hr of cultivation. Due to high CAT activities in undiluted extracts, dilutions were made (1 : 10, 1 : 100, and I : 1000) to obtain CAT activities in a linear range for densitometric scanning of autoradiographs. The results -~ W. Werr, W.-B. Frommer, C. Maas, and P. Starlinger, EMBO J. 5, 1373 (1985). 36 R. Liihrs and H. L6rz, Planta 175, 71 (1988). 37 C. Maas and W. Werr, Plant Cell Rep. 8, 148 (1989). ~8 C. Maas, S. Schaal, and W. Werr, EMBO J. 11, 3447 (1990).
76
VECTORS FOR EXPRESSINGCLONED GENES
•
A
HindllI
/
\ LSphI
(
~PstI
p R T 101cat
\\"
/ / BamHI
Hind III \ sphi ~ Pst I ~ H~ncI
/ / /~ / ..~.~Y
~
Xhol/
\ 4 42
II
pRT-ex/s-cat
Sma I
43
1084
I
~'
I
Sst I HindIII 4 42/43 []
~
tel. CAT activity 22 74
pRT-int/s-cat,
Barn HI XbaI 1084 1
!
Sst 1 HindIII
B
[6]
pRT-ex/s-int/s-cat
Barn HI XbaI
tobacco ex+int
(-)
barley 940
ex+int int
130
ex (-)
9 1 controls
no DNA CM only CAT enzyme 1,3
3
1
CM
[6]
GENE EXPRESSION IN DICOTS AND MONOCOTS
77
of the 1/10 dilution (equal to 1 × 105 protoplasts) of the extracts (adjusted to give comparable amounts of protein) are shown in Fig. 5B. Comparison of pRTlOlcat expression in protoplasts of the monocot barley and the dicot tobacco clearly reveals the reduced activity of the CaMV 35S RNA promoter in monocotyledonous plants (Fig. 5B). Expression driven from this promoter in barley was enhanced up to 940-fold by the combined action of a Sh 1 exon 1 and an intron 1 segment inserted into the transcription unit (pRT-ex/s-int/s-cat). Individual stimulation of gene expression by either the Sh 1 exon 1 (pRT-ex/s-cat) or the Sh 1 intron 1 (pRT-int/s-cat) is also observed. These results compare quite well with the data obtained by using protoplasts of other monocotyledonous plant species: Oryza sativa, Panicum maximum, Pennisetum purpureum, and Zea mays. 12,14 Data obtained from these experiments imply that the ninefold stimulation by Sh 1 exon 1, which is conserved among angiosperms, must occur at the transcriptional level. A position just downstream of the transcription start site is highly unusual for an RNA polymerase II-dependent transcriptional activator. However, reports concerning the human glial fibrilliary acidic protein 39 and the adenovirus IVa2 gene 4° have revealed other members of this new class of RNA polymerase II-dependent transcriptional activators. The dramatic 130-fold stimulation of gene expression by the Shrunken 1 intron 1 appears to be closely linked to the splicing process. Stimulation 39 y . Nakatani, M. Horikoshi, M. Brenner, T. Yamamoto, F. Besnard, R. Roeder, and E. Freese, Nature (London) 348, 87 (1990). 4o j. Carcamo, E. Maldonado, P. Cortes, M.-H. Ahn, I. Ha, Y. Kasai, J. Flint, and D. Reinberg, Genes Dev. 4, 1611 (1990).
FIG. 5. (A) Chimeric gene constructions used in transient gene expression experiments. From the maize Shrunken 1 (Sh 1) gene the Sh 1 exon (represented by the black boxes) and/or the Sh 1 intron 1 (represented by the open boxes, "intron 1") were inserted into the Sinai site of plasmid pRTlOlcat. The position of the terminal nucleotides of both Sh 1 elements is indicated. Restriction sites relevant for cloning have been included, ex, exon 1; int, intron 1 ; s, sense orientation of the insert. (B) Result of a CAT assay performed with a 1 : 10 dilution of crude extracts of tobacco and barley protoplasts that were homogenized 40 hr after DNA delivery, Relative CAT activity of barley protoplasts transfected with pRTlOlcat (--) is referred to 1. The constructions used have been indicated in abbreviated form beside the corresponding lane: (--), pRTlOlcat; ex, pRT-ex/s-cat; int, pRT-int/s-cat; ex + int, pRT-ex/s-int/s-cat. Autoradiographs were scanned densitometrically. CM, [~4Clchloramphenicol; CAT enzyme, chloramphenicol acetyltransferase enzyme; CM only, only [t4C]chloramphenicol was loaded; no DNA, protein extract of untransfected protoplasts was used; 1, 3, 3, 1, CM, acetylation products of the CAT reaction.
78
VECTORS FOR EXPRESSING CLONED GENES
[6]
can be observed only when intron 1 is positioned in the transcription unit and in the sense orientation. 12Data obtained from animal model systems imply that the molecular nature of stimulation by insertion of intron sequences is mainly due to the formation of spliceosomes and an increased resistance of the primary transcript to turnover processes by nuclease attack in the nucleus. 41 The failure of exon 1/intron 1 to stimulate gene expression in tobacco (pRT-ex/s-int/s-cat shows a 70% reduced value as compared to pRTlOlcat) could be due to apparent differences in the splicing process (e.g., formation of spliceosomes by correct splice site recognition) between monocotyledonous and dicotyledonous cells. 42'43 In conclusion, improvement of the cassette pRTlOlcat to produce pRT-ex/s-int/s-cat gave us a construct leading to a high level of gene expression in monocotyledonous plants and in particular in agronomically important cereals. The construct pRT-ex/s-int/s-cat has been constructed to facilitate the replacement of the CAT-coding region as a BamHI or XbaI fragment or an exchange of the promoter region as a HinclI/XhoI or HinclI/KpnI fragment. The entire chimeric gene can be transferred to other vectors using PstI or SphI. Moreover, the construct pRT-ex/s-int/scat might be helpful for the analysis of plant promoters or to achieve high levels of gene product in transgenic plants. Acknowledgment A part of this work was supported financially by Hoechst AG Frankfurt am Main, Germany.
41 A. R. Buchman and P. Berg, Mol. Cell Biol. 8, 4395 (1988). 42 B. Keith and N.-H. Chua, EMBO J. 5, 2419 (1986). 43 G. J. Godall and W. Filipowicz, Cell 30, 763 (1989).
[7]
OVERPRODUCTION OF PROTEINS USING ECPCR
79
[7] O v e r p r o d u c t i o n o f P r o t e i n s U s i n g E x p r e s s i o n - C a s s e t t e Polymerase Chain Reaction By K U R T I S D . M A C F E R R I N , L I N C H E N , MICHAEL P. TERRANOVA, STUART L . SCHREIBER, and GREGORY L . VERDINE
Introduction
Decades of basic research into the genetics of the enteric bacterium Escherichia coli 1 have yielded an outcome of immense practical value: a
working knowledge of gene architectures that afford high-level protein biosynthesis in this organism, and the means by which to construct them. This ability to rationally "engineer" protein-overproducing strains of E. coli has revolutionized protein science by making large quantities of recombinant proteins routinely available for study. Nonetheless, until recently the construction of overproducing E. coli strains was an often difficult and time-consuming task, primarily because of two factors: (1) the requirement that the gene or cDNA to be expressed be refashioned from a molecularly monoclonal source (such as a homgeneous plasmid, as opposed to a polyclonal source such as a library), and (2) the requirement to precisely assemble specific DNA sequences during construction of the recombinant (overexpressing) gene. We 2 and others 3-5 recognized the potential of the polymerase chain reaction 6 to overcome both these problems simultaneously by rapidly and selectively amplifying a target sequence from a mixed population, while at the same time allowing defined sequence alterations to be made in the amplification products. 7-9 In this chapter we briefly review the architecture of a gene construct suitabl~ for high-level, regulated expression in E. coli (henceforth called an overprot F. C. Neidhardt, "'Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology" (J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger, eds.), Vols. 1 and 2. American Society for Microbiology, Washington, D.C., 1987. 2 K. D. MacFerrin, M. P. Terranova, S. L. Schreiber, and G. L. Verdine, Proc. Natl. Acad. Sci. U.S.A. 87, 1937 (1990). 3 C. M. Skoglund, H. O. Smith, and S. Chandrasegaran, Gene 88, 1 (1990). 4 j. V. Gray, B. GolineUi-Pimpaneau, and J. R. Knowles, Biochemistry 29, 376 (1990). 5 X. R. Gu and D. V. Santi, DNA Cell Biol. 9, 273 (1990). 6 R. K. Saiki, S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. Arnheim, Science 230, 1350 (1985). v W. Bloch, Biochemistry 30~ 2735 (1991). 8 K. B. Mullis and F. A. Faloona, this series, Vol. 155, p. 335. 9 H, A. Erlich, D. Gelfand, and J. J. Sninsky, Science 252, 1643 (1991).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
80
VECTORS FOR EXPRESSING CLONED GENES
[7l
ducer) and discuss in detail our polymerase chain reaction (PCR)-based method for making overproducers--the expression-cassette polymerase chain reaction (ECPCR). 2 A r c h i t e c t u r e of O v e r p r o d u c e r T h e g e n e r a l a r c h i t e c t u r e o f a n o v e r p r o d u c e r is s h o w n in Fig. 1. Virtually all s u c h c o n s t r u c t s p o s s e s s as b a c k b o n e a p l a s m i d v e c t o r c o n t a i n i n g a s e l e c t a b l e p h e n o t y p i c m a r k e r (such as r e s i s t a n c e to ampicillin) a n d a
promoter
ribosome binding site (RBS)
coding sequence
vector ~ / / / ' J
....-
.I
L..
[ ....
transcription terminator L~--~X,~ vector
overproducer
. . . . . . . . :--:-,.,____ .....................
[ -35 transcriptional -10 " ' ] ""] /sequence spacer sequence I Shine- translational| start E. coil ...TTGACANNNNNNNNNNNNNNNNNTATAAT... Dalgamo spacer / codon nominalp r o m o t e r ...AGGAGGNNNNNNNNATG... T7 ...GAAATTAATACGACTCACTATAGGGAGA...
I
II transcriptional L . . . . . . . . . . -,,.control
I
transcription translational control
...A GGA G GNNNNNNNNA UG.. "1
mRNA
I translation -.=-. . . . . . .
J
ro,e, I nl FIG. 1. Architecture of an overproducer. At the top is a schematic diagram of the sequence elements that are present in an overproducer; in the expansions below are subelements and actual sequences that yield high-level expression in E. coll. The promoter and ribosomebindingsite (RBS) control the initiationof transcription and translation, respectively. Promoters recognized by both the E. coil and bacteriophage T7 RNA polymerase are shown; either promoter can function with the same RBS sequences. For simplicity, the start and stop codons are not depicted; in subsequent figures, they are shown explicitly wherever necessary. All sequences correspond to the coding strand, written in the 5' ~ 3' direction reading from left to right.
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
81
high-copy-number origin of replication (not shown). Fused directly to the upstream end of the coding sequence (start codon, not shown) is the ribosome-binding site or RBS, which functions in the transcript to control the efficiency of translation initiation, w The RBS is composed of two subelements: (1) the Shine-Dalgarno sequence, which is believed to basepair with the 16S subunit of the ribosome, and (2) the translational spacer, which positions the start codon on the ribosome relative to the anchored Shine-Dalgarno sequence. In the transcript, formation of secondary structure involving the RBS apparently reduces its efficiency, l° and because the flanking gene sequence is different in each construct, it is perhaps unsurprising that no universally optimal RBS has been discovered. Nonetheless, the following general rules can be applied to the design of RBSs: (1) the Shine-Dalgarno sequence should contain all or part of the sequence 5'-AGGAGG-3'--some evidence suggests that longer sequences are preferred; and (2) although there appears to be no strict sequence requirement for the translational spacer, the most efficient appear to possess four to eight nucleotides having an adenine/thymine-rich composition. ~° Fortunately, several RBSs have been discovered that seem to work well with most coding sequences (see below). Upstream of the RBS [often - 1 0 - 1 0 0 base pairs (bp)] is positioned the promoter, which controls the initiation of transcription by binding directly to a DNA-dependent RNA polymerase. 1~ Promoters that are recognized by E. coli RNA polymerase 12 and bacteriophage T7 RNA polymerase 13 are in common use today (Fig. 1). The consensus E. coli promoter has the following physical organization: (1) the hexanucleotide sequence 5'TTGACA-3', centered approximately 35 bp upstream of the transcription start point (and thus designated the - 3 5 sequence), followed by (2) a 16to 17-bp sequence-independent spacer, followed by (3) the hexamer 5'TATAAT-3' ( - 10 sequence). The most efficient T7 promoter, on the other hand, consists of a 17-bp nominal promoter (positions - 1 to - 17) flanked on each end by additional defined sequences that contribute modestly to the efficiency of transcription initiation. 13-~5 Promoters used for protein overproduction should not only be strong, but should also be tightly regulated. Expression systems controlled by the E. coli lac repressor, espel0 L. Gold, Annu. Rev. Biochem. 57, 199 (1988). Ii p. H. von Hippel, D. G. Bear, W. D. Morgan, and J. A. McSwiggen, Annu. Rev. Biochem. 53, 389 (1984). t2 W. R. McClure, Annu. Rev. Biochem. 54, 171 (1985). 13 W. F. Studier and J. J. Dunn, J. Mol. Biol. 166, 477 (1983). 14 Z. D. Schneider and G. D. Stormo, Nucleic Acids Res. 17, 659 (1989). 15 E. D. Jorgensen, R. K. Durbin, S. S. Risman, and W. T. McAllister, J. Biol. Chem. 266, 645 (1991).
82
VECTORS FOR EXPRESSING CLONED GENES
[7]
cially those having the lacUV5,16,17 tac, 18,19and trc 2° promoters, have been most widely used, because they can be induced simply by the addition of isopropyl-fl-o-thiogalactopyranoside (IPTG) to the medium. Another highly successful system utilizes the bacteriophage k promoter P L in combination with the temperature-sensitive k repressor mutant cI85721'22; these can be induced thermally by shifting the temperature of the growth medium from - 3 0 to 42° . The features of these and other systems have been discussed in detail elsewhere. 23,24 Finally, downstream of the coding sequence (with no precise distance requirements) most overproducers possess a transcription terminator,~l,~2 which forms a stem-loop structure in the transcript and thereby aborts transcription, independent of the factor rho. This element enhances the economy of nucleotide utilization and increases plasmid stability by preventing unnecessary transcription beyond the 3' boundary of the recombinant gene. Because many sequences are capable of functioning as transcription terminators, and these elements are generally included as part of the overexpression vector, they are not discussed further here. In summary, the promoter and transcription terminator are transcriptional control elements, they do not have to be positioned precisely with respect to the coding sequence, and their efficiency is essentially independent of the coding sequence; the RBS, on the other hand, is a translational control element, it must be positioned precisely with respect to the start codon, and varies in efficiency (for a given RBS sequence) with different 5'-coding sequences. Finally, construction of an overproducer involves the assembly of these expression and coding elements in the required configuration.
Overproducer Construction Using Expression-Cassette Polymerase Chain Reaction In assessing various general strategies for overproducer construction, we were attracted by the concept of treating the RBS and coding sequence as linked units, because they exert a coordinate influence over translation 16 C. Yanisch-Perron, J. Vieira, and J. Messing, Gene 33, 103 (1985). 17 F. W. Studier and B. A. Moffat, J. Mol. Biol. 189, 113 (1986). ~s H. de Boer, L. J. Comstock, and M. Vasser, Proc. Natl. Acad. Sci. U.S.A. 80, 21 (1983). 19 E. Amman, J. Brosius, and M, Ptashne, Gene 25, 167 (1983). 2o E. Amman and J. Brosius, Gene 40, 183 (1985). 21 S. Tabor and C. C. Richardson, Proc. Natl. Acad. Sci. U.S.A. 82, 1074 (1985). 22 W. Whalen, B. Ghosh, and A. Das, Proc. Natl. Acad. Sci. U.S.A. 85, 2494 (1988). 23 A. Das, this series, Vol. 182, p. 93. 24 D. V. Goeddel (ed.), this series, Vol, 185.
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
R. site
L_
expression cassette: transcription vector:
stop codon
start codon
RBSI
I~//,/,~:. I
83
coding sequence
I i l l i . - a. site d
promoter ~ t .
term.
.
site(s) promoter
start codon RBS I R. site
X'R"
overproducer
stop codon I t. term. R, site
FIG. 2. A strategy for overproducer construction that treats the coding sequence and RBS as linked units. In this scheme, an expression cassette containing all of the required translational information is inserted into a vector containing all of the required transcriptional information (transcription vector). R. site, Restriction endonuclease cleavage site; t. term, transcription terminator.
initiation. Implementation of this strategy in overproducer construction (Fig. 2) would require assembly of an expression cassette containing the RBS and coding sequence, and then insertion into a vector containing the necessary transcriptional control elements (transcription vector). Expression could then be optimized by changing the sequence of the RBS within the expression cassette, without having to alter the sequence of the vector. We also wanted to be able to package readily any contiguous coding sequence into the expression cassette, regardless of where it was found within the wild-type coding sequence (donor); this would make it possible to overproduce predetermined segments of a gene corresponding to, for example, individual domains of a protein. Although the construction of such expression cassettes has been accomplished by conventional recombinant DNA methodology, we wanted to avoid the demands of time and specialized training associated with such methods. Finally, we wanted to be able to synthesize expression cassettes directly from donors that are present in libraries, so as to avoid having to reclone known genes or to obtain molecular clones from elsewhere. The method that has emerged from these design considerations, ECPCR, fulfills the foregone objectives by using PCR in the synthesis 25,26 of expression cassettes. 25 K. Mullis, F. Faloona, S. Scharf, R. Saiki, G. Horn, and H. Erlich, Cold Spring Harbor Symp. Quant. Biol. 51, 263 (1986). 2~ S. J. Scharf, G. T. Horn, and H. A. Erlich, Science 233, 1076 (1986).
84
VECTORS FOR EXPRESSING C L O N E D GENES
[7]
A schematic diagram of ECPCR is presented in Fig. 3. Two synthetic oligonucleotide are designed to equip a coding sequence with translational control information and restriction sites during PCR (Fig. 3A); use of these primers in PCR furnishes an expression cassette that contains all of the information necessary for cloning and translation in E. coli (Fig. 3B). Specifically, the 5' primer contains an N-terminal coding sequence for the desired recombinant protein, preceded by a start codon (which may or may not be present in the donor), translational spacer, Shine-Dalgarno sequence, restriction site, and an end clamp (which facilitates cleavage by the restriction enzyme). The 3' primer contains a C terminal anticoding sequence for the desired recombinant protein, preceded by a stop anticodon, restriction site, and end clamp. Polymerase chain reaction amplification of the donor using 5' and 3' primers yields the desired expression cassette, which can be cloned into a transcription vector either directly or after restriction cleavage (cf. Fig. 2). More detailed aspects of primer design, sources of donor DNA, and experimental procedures are provided below. In our first reported example, ECPCR was used to construct overproducers encoding specific fragments of CD4, the cell surface receptor for major histocompatibility complex class II antigens (complexed to the T cell receptor and peptide antigens) and the human immunodeficiency virus (HIV) surface glycoprotein gpl20. 2 Based on the known domain organization of the immunoglobulin-like CD4, we designed two 5' primers that respectively encoded the N termini of CD4 domains 1 and 2, and 3' primers that encoded the C termini of domains 1 and 2, fused to the required translational control and cloning sequences. Polymerase chain reaction using these primers in specific combinations (and a human CD4 donor) yielded expression cassettes that encoded CD4 domain 1, domain 2, and domains 1 + 2 (Fig, 4). On insertion into a transcription vector (see below), these expression cassettes directed high-level biosynthesis of the respective CD4 fragments, and these are now being used in ongoing studies to assess their individual contributions to overall receptor binding--this approach we have termed "domainal analysis. ''2 Design of Primers 5' Primer
In our original ECPCR study, 5' primers [Fig. 3A (a)] were used, with coding sequence corresponding to CD4 residues Lys ~ to Asp ~° (putative N terminus of domain 1) and Asn ~°3 to Gin lj2 (putative N terminus of domain 2), in addition to the initiator methionine codon. The translational
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
A
85
5'-Drimer" RBS
(a) (b) (c) (d)
end restrictionr Shine- translationa~ start codingsequence clamp sit~ Dalgarno spacer codon (N-terminus) 5'-CGCGCGAATTCAGGAGGAA'FI'TAAAATG-(30 nucleotides)-3' 5'-TAGGGCGAATTCAAGGAGATATACCATG-(26 nucleotides)-3' 5'-TAGGGCGAATTCAAGGAGATATACATATG-(26 nucleotides)-3' 5'-TAGGGCGAATTCTTAACCAGGGAGCTGA'I-rATG-(26 n'tides)-3'
[
I
J
.I
Donor:
l
n
ni i i i i i i
,
r .............. (e) (f)
protei"-cedingDNA
Exoresslon
. . . . . . . . . . --.L'.'.---:'..J ..J . J
r ..... s = " ; = /
/
...... ; ....... /
PCR
cassette
5' restrictionsite end clamp
t~_~ _~ ~J~ !
3'-(30 nucleotides)-ATTTTCGAAACGCG-5' 3'-(26 nucleotides)-ATT TTCGAAACGGATCT-5' anticodingsequence stop restriction end (C-terminus) anticodon site clamp
~ B
of coding strancl
complementary region of anticoding strand
3'-orimer:
.J
:::::::::::::::::::::: ............ com,,.men,ary,eg,on
"i t'"r ...... . ....... . ~ [-r ..... f: ........ ~
.J
3' restrictionsite
RiS r'-- start co(Ion
)
I I derived from 5'-primer
derived from donor
stopcodon
)
1
end clamp
I I derived from 3'-primer
FIG. 3. Schematic diagram of ECPCR. (A) D N A components used in ECPCR, and (B) structure of an ECPCR expression cassette. In ECPCR, a 5'-primer and a 3'-primer are used to simultaneously amplify a target coding sequence from a donor and to equip that sequence with translational control and cloning information. The Y-primer (a-d) contains coding sequence for the N terminus of the desired protein, in addition to a start codon, RBS, restriction site, and end clamp. The 3'-primer (e and f) contains anticoding sequence for the C terminus of the desired protein, in addition to a stop anticodon, a restriction site, and end clamp, Use of a 5'- and 3'-primer to amplify a donor coding sequence results in the synthesis of an expression cassette in which parts of the coding sequence are primer derived and parts are donor derived.
86
[7]
VECTORS FOR EXPRESSING C L O N E D G E N E S
(A)
{B)
(C)
tn
CD4 SM
cDNA
1353 603 310 \ 281/271 234/--"
Primers
Start-1
+
+
-
Start-2 Halt-1
+
+
-
+ -
+
+
Halt-2
-
+
+
+
+
+
+
+
+
+
+
+
-
-
FIG. 4. PCR synthesis of CD4 expression cassettes using various donors. 2(A) CD4 cDNA; (B) Human KBM-7 myeloblastoid cell cDNA library, human HPB-ALL T cell cDNA library, and total Jurkat T cell RNA; (C) BALB/c mouse mRNA from various tissues. The Start-1 and Start-2 primers correspond to primer (a) of Fig. 3A, with coding sequence for the N termini of CD4 domains 1 (Lys I to Asp j°) and 2 (Asn 1°3to Glnm), respectively; the Halt-1 and Halt-2 primers correspond to primer (e) of Fig. 3A, with anticoding sequence corresponding to the C termini of CD4 domains I (Va198to His I°7)and 2 (VaP 68to Leut77), respectively. " R N A " samples are actually R N A - D N A hybrids generated from first-strand cDNA synthesis. Lane SM, DNA size markers (HaeIII-digested 4~X174), with fragment sizes indicated in bp. In (B) only the products from amplification using the Start-2 and Halt-2 primers (domain 2 expression cassettes) are shown; the other two primer combinations give rise to their respective expression cassettes. In (C) the products from amplification using the Start-1 and Halt-1 primers (domain 1 expression cassettes) are shown. The expression cassettes in (C) encode murine-human chimeras, with the primer-derived sequences being human and the donorderived sequences being murine (cf. Fig. 3B). Further details are provided in Ref. 2.
[7l
OVERPRODUCTION
OF PROTEINS USING E C P C R
Sma I Xma I EcoR I
87
Acc I Sal I BamH I
Pst I
Hind III
I ...GAATTCCCGGGGATCCGTCGACCTGCAGCCAAGCTT...I
rminalor
(ampicillin
resistance
"
transcription vector
gene)
FIG. 5. Physical map and polylinker sequence of the transcription vector pHN 1 +.
spacer element, 5'-AATTTAAA-3', was designed to possess 8 bp of A/Trich sequence plus a low-frequency restriction site (DraI, 5'-TTT $ AAA3'), which would allow for future insertion of a periplasmic signal sequence. The consensus Shine-Dalgarno, 5'-AGGAGG-3', was used. The EcoRI site, 5'-GAATTC-3', was chosen for use in the 5' primer, because in the transcription vector we use (pHN 1 +) (Fig. 5; also see below) the EcoRI site is closest to the promoter; this site was flanked by a 5-bp GC clamp, according to the suggestion of Scharf. 27 Since that report, we have tested a number of additional 5' primers, the best of which are shown in Fig. 3A (b-d). Primers (b) and (c) contain variants of an RBS sequence from T7 gene 10, which have seen popular use in transcription/translation vectors 2~'24; the use of the primer (c) trans27 S. J. Scharf, PCR Protocols 1, 84 (1990).
88
VECTORS FOR EXPRESSING CLONED GENES
[7]
lational spacer for ECPCR was first reported by Liu et al. 28 Primer (d) contains the RBS from the E. coli ada gene, which we have also found to be highly efficient in several cases. 29 Although the RBSs of primers (b-d) are all strong, we have observed dramatic variations in their individual efficiencies when linked to the same gene, with no consistent trend among the RBSs when comparing them linked to different genes. 29 These results underscore the advantage of screening several RBSs as a strategy for maximizing expression. With regard to the restriction site in the 5' primer, we have almost exclusively used the E c o R I site, because it is the most promoter-proximal site available in pHN 1 + (Fig. 5). In one case we have observed a marked decrease in expression level when changing from a 5'-EcoRI expression cassette to an otherwise identical B a m H I cassette [made using a 5' primer, (d) in Fig. 3A, in which the restriction site was 5'-GGATCC-3'], each cloned into the corresponding polylinker site of pHN1 +.3° Secondary structure analysis 31 on the two transcripts has suggested that the mRNA of the B a m H I construct is able to form a stable hairpin that renders its Shine-Dalgarno sequence unavailable for pairing with the ribosome, and that this problem might be overcome by changing the RBS. At the design stage, there are perhaps two reasons why one would need to change the restriction site from E c o R I to another: (1) if the cleaved E c o R I site (5'AATT-3' overhang) is not compatible with sites available in the transcription vector, or (2) if there is an E c o R I site within the coding sequence to be included in the expression cassette. Because cleavage of a restriction site near an end is much slower than at an internal site, 32one can effectively rule out using partial digestion to cut a primer-derived site in an expression cassette without affecting an internal site. We have sometimes experienced considerable difficulty with restriction digestion of expression cassettes having GC clamps [as in Fig. 3, primers (a) and (e)], and so now avoid their use altogether. The end-clamp sequences in primers (b-d) were chosen because they flank the E c o R I site in common polylinker sites, and are thus known not to inhibit cleavage; they have performed well without exception in our hands. In general, we prefer to use end clamps of 5 bp or more--even though reports indicate that shorter sequences are efficiently cleaved by some restriction endonu28 j. Liu, M. W. Albers, C. M. Chen, S. L. Schreiber, and C. T. Walsh, Proc. Natl. Acad. Sci. U.S.A. 87, 2304 (1990). 29 L. Chen, W. Chang, K. Ezaz-Nikpay, C. J. Larson, M. P. Terranova, and G. L. Verdine, unpublished observations. 3o L. Myers and G. L. Verdine, unpublished observations. 3~ D. E. Cane, personal communication. 32 W. E. Jack, B. J. Terry, and P. Modrich, Proc. Natl. Acad. Sci. U.S.A. 79, 4010 (1982).
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
89
cleases33,34--because the relative cost of tracking down a recalcitrant cleavage site far outweighs that of adding a few base pairs to the primer. Design considerations that apply to the coding sequences in the 5' primer are discussed below.
3'-Primer In our original ECPCR study, 3' primers [(e) in Fig. 3] were used, with anticoding sequence corresponding to CD4 residues Va198 to His ~°7 (putative C terminus of domain 1) and Va1168 to L e u 177 (putative C terminus of domain 2), fused to a stop anticodon, HindlII site, and GC clamp (see also Fig. 4). The design of the 3' primer has since changed only in that we have replaced GC clamps with mixed-sequence end clamps [Fig. 3A, primer (f)]. There are three possibilities for the stop anticodon, namely 5'-TTA-3', CTA, and TCA, and all should work equally well except in suppressor (sup) E. coli strains, which may read through one of the three. 35 The choice of restriction site is dictated by those available in the transcription vector and absent from the coding sequence. We chose HindlII because it is the polylinker site farthest from EcoRI in pHN1 + (Fig. 5).
Coding and Anticoding Sequences in Expression-Cassette Polymerase Chain Reaction Primers Many of the design considerations that apply to the donor-homologous regions of ECPCR primers apply in general to primer design in most PCRbased methods, and this subject has received a great deal of attention. In the domainal analysis of CD4, 2 we used 30-bp donor-homologous sequences in the 3' and 5' primers [Fig. 3A, primers (a) and (e); Fig. 4]. We have since found that sequences of this length are generally not necessary, and in fact longer sequences are sometimes undesirable because of their greater probability to participate in PCR side reactions. If a primer longer than a 60-mer must be used, programs such as Fold and Squiggles (GCG) are useful in determining the stability of potential secondary structures, which may be reduced through alternative primer formulations. We have found that 18 bp of donor-homologous sequence is generally sufficient,29 33 D. L. Kaufman and G. A. Evans, BioTechniques 9, 304 (1990). 34 New England Biolabs 1990-1991 catalog, p. 132. 35 B. Lewin, "Genes IV," pp. 150-153. Oxford Univ. Press, Oxford, and Cell Press, Cambridge, England, 1990.
90
VECTORS FOR EXPRESSING CLONED GENES
[7]
especially if one is using a monoclonal donor. An overview of principles for primer design is given by Innis and Gelfand. 36 Materials and Methods
Oligonucleotides All oligonucleotides were synthesized on a 381A automated DNA synthesis machine (Applied Biosystems, Foster City, CA) on a 0.2-~mol scale, using/~-cyanoethylphosphoramidites. Oligonucleotides were either purified Trityl-on using OP cartridges (Applied Biosystems) or, for Trityloff oligonucleotides, by acrylamide gel electrophoresis as follows37: following ammonia deprotection and lyophilization, the solid oligonucleotide was dissolved in 200/~1 formamide loading dye, heated to 90° for 2 min, and loaded onto a 20% (w/v) denaturing acrylamide gel (20 × 20 × 0.3 cm; 19 : I acrylamide : methylenebisacrylamide; 7 M urea). The gel was run until the bromphenol blue marker migrated out of the gel, then the gel was removed, wrapped in Saran (plastic wrap), and placed on top of a fluorescent-backed thin-layer chromatography (TLC) plate. The gel was then visualized by illumination with short-wave ultraviolet (UV) light, and the uppermost band was cut out with a sharp scalpel. The gel slices were transferred to a 15-ml Corning tube and crushed thoroughly with a glass rod. Following the addition of 10 ml 1 M aqueous triethylammonium bicarbonate (TEAB) (made by adding solid CO2 to a triethylamine solution until the pH reaches 8.0), the tube was shaken moderately for 8-12 hr at room temperature. The solution was removed, being careful to avoid the gel bits (higher recoveries can be obtained if the gel slices are washed with 10 ml of 25 m M TEAB), and transferred to a syringe fitted with a SepPak C1s cartridge (Waters Associates, Milford, MA); the SepPak had previously been washed with the following: 20 ml of CHaCN, 10 ml of 30% (v/v) CH3CN in 0.1 M TEAB, and 20 ml of 25 mM TEAB. The oligonucleotidecontaining solution was passed twice through the SepPak at approximately 1 drop/sec (this can be done conveniently using a syringe pump), followed by a wash of 10 ml 25 mM TEAB. The oligonucleotide was then eluted using 5 ml of 30% (v/v) CH3CN in 0.1 M TEAB and lyophilized (SpeedVac; Savant Instruments, Farmingdale, NY) to dryness. Dry oligonucleotides from cartridge or gel purification were redissolved in 100/xl TE buffer (10 mM Tris-HC1, 1 mM EDTA, pH 8.0) and were quantified by measuring the absorbance at 260 nm of 1 /.~1ofoligonucleotide solution diluted in 1 ml 36 M. A. Innis and D. H. Gelfand, PCR Protocols 1, 3 (1990). 37 T. Ellenberger, personal communication.
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
91
TE buffer. The absorbance was multiplied by 64,000 and divided by the number of bases in the oligonucleotide to provide a rough estimate of concentration in units ofpicomoles per microliter. We have found cartridge purification to be sufficient in most cases, especially for <30-mer oligonucleotides. The gel method is preferred for >30-mers, and is somewhat more convenient for large-scale purifications (greater than 200 nmol).
Template DNA We have successfully generated expression cassettes using the following donors: minipreparation monoclonal plasmids, a hgt 10 library, a phagemid library (minipreparation), a first-strand cDNA library ]generated from total RNA primed with poly(dT) and extended with reverse transcriptase--the R N A - D N A hybrid mixtures were used directly], and total E. coli DNA. In the case of CD4, we were able to generate mouse/human chimeras in which the primer-encoded sequences were human and the donor-encoded sequences murine (see Fig. 3, bottom; Fig. 4C) by using human primers to amplify murine first-strand cDNA libraries." (The success of such chimera experiments depends on the number and location of mismatches.) While we take no particular care to purify the donor, it is a general rule in PCR that purer is better.
Overexpression System We generally use the transcription vector pHN 1 + 38.39(Fig. 5), a phagemid in which cloned genes are under the control of the tac promoter 18'~9 and rrnBT~T 2 transcription terminator. 4° Owing to the high copy number of p H N I + constructs, they are best maintained in E. coli strains that express a high titer of the lac repressor, for example XA90 (F' laclQIproAB) [~lac pro XIII ara nal argE(am) thi rifr]. Other transcription plasmids such as pKK223-34~ (tac promoter, rrnBT~T2 transcription terminator) and some members of the pET series 24(T7 promoter and terminator; Novagen, Madison, WI) are equally suitable.
Enzymes~Reagents Thermus aquaticus (Taq) DNA polymerase was obtained from Promega (Madison, WI). The Sequenase kit (U.S. Biochemical, Cleveland, 38 L. C. Myers, M. P. Terranova, H. M. Nash, M. A. Markus, and G. L. Verdine, Biochemistry 31, 4541 (1992). 39 S. L. Schreiber and G. L. Verdine, Tetrahedron 47, 2543 (1991). 4o j. Brosius, T. J. Dull, D. D. Sleeter, and H. F. Noller, J. Mol. Biol. 148, 107 (1981). 4t j. Brosius and A. Holy, Proc. Natl. Acad. Sci. U.S.A. 81, 6929 (1984).
92
VECTORS FOR EXPRESSING CLONED GENES
[7]
OH) was used for all DNA sequencing. All other enzymes were obtained from Boehringer Mannheim (Indianapolis, IN) or New England BioLabs (Beverly, MA). Deoxynucleotide triphosphates were obtained from Pharmacia (Piscataway, N J). Perfect Match was obtained from Stratagene (La Jolla, CA).
Thermal Cycler A programmable thermal controller (MJ Research, Cambridge, MA) was used for all amplifications.
Software The University of Wisconsin Genetics Computer Group (Madison, WI) version 7.0 suite (GCG) 42of nucleic acid and protein analysis software was accessed on a VAX 8700 running VMS 5.4-2.
Amplification There are a large number of published protocols for PCR, and the optimal conditions seem to vary somewhat with different donor and primer preparations. The optimization of PCR has been well discussed elsewhere, 36 but the parameters that in our experience are most likely to require adjustment are donor and primer concentration, annealing temperature, and number of cycles. In general, we have found the procedure used to produce CD4 expression cassettes 2 to be a good starting point: in a sterile 0.5-ml microcentrifuge tube are combined 1/zl of a 40 pM donor solution (for monoclonal donors), 5/.d of a solution containing 10/zM concentrations of each primer, 5/xl of a solution containing 2.5 mM concentration of each dNTP (N = A, C, G, T), 0.5/zl Taq polymerase (5 U/t~I), and 5/~1 of 10 x Taq buffer [supplied by Promega: 500 mM TrisHC1 (pH 9.0 at 25°), 500 mM NaC1, 100 mM MgCI2, I% (w/v) Triton X-100]. Then, 28.5/xl of sterile double-distilled water is added to produce a final volume of 50 p,l. The capped tube is shaken and microcentrifuged briefly, and overlaid with mineral oil until the aqueous layer is fully covered (-25/zl). The tube is placed in a temperature cycler and heated at 94 ° for 1 min. The following cycle is then repeated 30 times: heat at 94° for 30 sec; 42 °, 1 rain; 70 °, 1 min. After this the reaction is held at 70 ° for 10 min to ensure complete extension. The reaction is then allowed to cool to room temperature and microcentrifuged briefly. The amplification procedure is essentially the same when using nonmonoclonal donor sources, such as cDNA libraries or crude first-strand 42 j. Devereux, P. Haeberli, and O. Smithies, Nucleic Acids Res. 12, 387 (1984).
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
93
cDNA preparations, although in this case it is worthwhile to carry out serial dilution of the donor. We have successfully used 1 /zl or less of a 10-ng/tzl cDNA library, and 1 /xl or less of a 1-ng//xl first-strand cDNA preparation as the source of donor DNA in ECPCR. An aliquot (5 /xl) of the crude PCR reaction should be analyzed by agarose gel electrophoresis using ethidium bromide staining (5/zl of a 10mg/ml solution per 100-ml gel solution) and visualized using short-wave UV light, to evaluate the success of the reaction (Fig. 4). Size markers should be used to aid in determining the size of the PCR products. Because unreacted primers can often be seen as blurred bands in the low molecular weight range, and template DNA may also be visible in some cases, it is worthwhile to include these samples as controls. It is also worthwhile to carry out a diagnostic restriction analysis of the PCR product before inserting it into a vector, so as to confirm that it is the expected product. The PCR reaction mixture is worked up as follows: after brief microcentrifugation, the reaction mixture is frozen by placing it in a - 20° freezer for a few minutes, then the mineral oil is removed with a pipette. TE buffer (50 tzl) is added, followed by an equal volume ( - 100 tzl) of phenol/ chloroform solution (equal volumes of phenol and chloroform saturated with 0.1 M Tris-HCl, pH 7.6), and the mixture is vortexed gently. The tube is microcentrifuged for 15 min, and the upper (aqueous) layer is removed and added to a new 1.5-ml microcentrifuge tube. The extraction is repeated once, and then the tube is spun for 1-2 min in a centrifugal lyophilizer (SpeedVac) to remove traces of chloroform. The volume is then adjusted to 500/zl with TE buffer, followed by the addition of 50/zl sodium acetate buffer (3.0 M, pH 6.0) and 1 ml of 100% ethanol chilled to - 20 °. The tube is vortexed briefly and chilled at - 20° for 30 min. (Sodium acetate should be used instead of ammonium acetate to avoid problems with inhibition of ligase by ammonium ion, and to more selectively precipitate DNA rather than oligonucleotides.) The tube is then microcentrifuged for 15 min, the solution is carefully decanted or removed by pipette, and the pellet (which may not be visible in some cases) is washed with 1 ml of 95% ethanol (room temperature). The tube is then left to air dry, after which 50/zl of TE buffer is added and repeatedly pipetted to effect dissolution. There are many adequate protocols for the remaining steps: digestion, ligation, and transformation. We favor the following protocols.
Digestion and Ligation The EcoRI and HindlII restriction sites of the CD4 expression cassettes 2 are cleaved in one digest as follows: in a 0.5-ml microcentrifuge tube 4/zl of the expression cassette in TE buffer (see above) is added to
94
VECTORS FOR EXPRESSING CLONED GENES
[7]
1/xl of a 1-mg/ml bovine serum albumin (BSA) solution, 1/xl of 0.1 M dithiothreitol (DTT) solution, and 1 /xl of 10 x buffer (500 mM NaCI, 100 mM MgCI2, 100 m M Tris-HC1, pH 7.7). Then 0.5/~1 EcoRI (8 units/p.l) and 0.5 ~1HindlII (12 units//.d) are added, followed by 2/xl doubly distilled water, to make a final volume of 10 ~1. The tube is vortexed gently and microcentrifuged briefly to ensure mixing of the reactants, then warmed to 37° for more than 1 hr. The tube is cooled to room temperature and spun briefly to collect the condensate. Agarose loading dye [2/xl: 50% (v/v) glycerol, 0.25% (v/v) xylene cyanole] is then added. Method A: Low-Melting-Point Agarose. 43 The entire sample of DNA in loading solution is loaded onto a 1% (w/v) SeaPlaque agarose (FMC, Rockland, ME) minigel (Hoefer, San Francisco, CA) in TBE buffer (90 mM Tris base, 90 mM boric acid, 2 mM EDTA). The gel and running TBE buffers are supplemented with ethidium bromide solution (1/.d of a 10 mg/ml ethanolic ethidium bromide stock solution per 100 ml gel solution and running buffer). Electrophoresis proceeds at 75 V for 1 hr, after which the gel is visualized with long-wavelength UV light, and the desired band is excised with a sharp scalpel. The gel slice is then trimmed with the scalpel to remove excess (nonfluorescing) gel. The trimmed gel slice (typically around 20/zl in volume) is added to a 0.5-ml microcentrifuge tube and diluted with an equal amount of sterile double-distilled water. The tube is heated to 70° for 5 min, shaken and microcentrifuged briefly, and placed in a 42 ° bath. A 5-/zl aliquot of the cleaved expression cassette in melted low-meltingpoint (LMP) agarose is transferred to a new 0.5-ml Eppendorf tube, preheated to 50 °. To this is added 5/xl of vector solution that had been purified similarly in LMP agarose (after restriction cleavage of 1/~g vector, followed by 5' dephosphorylation--it is convenient to purify the vector and the insert on the same gel). After mixing at 50°, 4/xl of 5 x T4 ligase buffer (10 m M DTT, 10 m M MgCI2, 0.6 m M ATP, 20 mM Tris-HCl, pH 7.6) and 5/xl sterile double-distilled water are added, and the tube is allowed to cool to room temperature. T4 DNA ligase (1 /zl at 1 U/~I) is added, the tube is shaken and microcentrifuged briefly, and the ligation mixture is then incubated at 16° overnight (about 12 hr). The ligation solution often resolidifies, but this does not prevent ligation. It is advisable to perform a control ligation in parallel, in which only vector and no insert DNA is present, to judge the background level of vector religation. Method B: Purification on Ceramic Beads, Ligation in Solution. The purification of the restriction-cleaved expression cassette is carried out using the GeneClean kit (BIO 101, La Jolla, CA) according to the proce43 L. S h e w c h u k , personal c o m m u n i c a t i o n .
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
95
dure provided by the manufacturer, with one additional step: after dissolution of the gel and addition of the ceramic beads, the mixture is incubated on ice for 5 min, shaking gently about once per minute to resuspend the beads. The DNA is eluted from the beads into 20/zl TE buffer, and this solution is used directly in ligation. A 5-tzl aliquot of the cleaved expression cassette in TE buffer is transferred to a new 0.5-ml Eppendorf tube. To this is added 5/xl of vector in TE buffer (which had been purified in parallel with the expression cassette, after restriction cleavage of 1/.~g vector, and 5' dephosphorylated), 10/xl 5 x T4 DNA ligase buffer (see above), 24/xl sterile, doubly distilled water, and 1 ~1 T4 DNA ligase (1 U//~I). The ligation is allowed to proceed overnight at 16°. It is advisable to perform a control ligation in parallel, in which only vector and no insert DNA is present, to judge the background level of vector religation.
Transformation Component E. coli XA90 is prepared and transformed by the method of Hanahan (protocol 1). 44 For DNA prepared by method A, the following changes are made: the ligation mixture is warmed to 42 ° for 5 rain (to melt any solid agarose), and 5/zl is added to the competent cells on ice, after which the transformation mixture is repeatedly and gently pipetted. After growth in SOB medium, 44 the transformed cells are spread on an LMA plate [1% (w/v) Bacto-tryptone, 0.5% (w/v) yeast extract, 1% (w/v) NaCI, 1.5% (w/v) Bacto-agar, 10 mM MgCIz, 0.1 mg/ml ampicillin-sodium salt], and the plate is incubated at 37° until colonies become visible (about 16 hr). Individual colonies are used to inoculate 3 ml sterile LBA medium [1% (w/v) Bacto-tryptone, 0.5% (w/v) yeast extract, 1% (w/v) sodium chloride, 0.1 mg/ml ampicillin-sodium salt], being sure to sample a range of colony sizes (overproducers often form smaller colonies than nonoverproducers). After being shaken 12 hr at 37°, the cells (1.5-ml inoculate) are harvested by microcentrifugation, plasmid DNA is isolated, 45 and the plasmids are checked for the presence of the insert by digestion with EcoRI plus HindIII, and insert-positive clones are sequenced. 46 Regarding the choice of method A vs. method B: we generally prefer method B, because it is less technically demanding and usually yields a
44 D. Hanahan, in " D N A Cloning" (D. M. Glover, ed.), Vol. 1, p. 119. IRL Press, Washington, D.C., 1985. 45 G. Del Sal, G. Manfioletti, and C. Schneider, BioTechniques 7, 514 (1989). 46 F. Sanger, S. Nicklen, and A. R. Coulson, Proc. Natl. Acad. Sci. U.S.A. 74, 5463 (1977).
96
VECTORS FOR EXPRESSING C L O N E D GENES
[7]
larger number of insert-positive transformants. However, with some difficult ligations, particularly those involving blunt ends, method A is sometimes superior to method B. By using method B in ligation of sticky-ended expression cassettes we typically obtain 100-200 colonies per plate, of which approximately 90% are the desired overproducer. 29 Induction Testing directly for induced overexpression of the desired protein can be used as an alternative to restriction screening of plasmids (but does not obviate the need for sequencing). If one has a convenient activity assay for the protein, it may be possible to assay crude cell lysates directly. Perhaps the most exciting method involves analysis of whole-cell protein by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE), in which case expression of the cassette-encoded protein can often be observed visually. SDS-PAGE also offers a convenient way to screen a series of expression cassettes that differ in their RBS sequences, which (as mentioned above) is an effective strategy for maximizing gene expression. We favor the following procedure for induction tests: single colonies from a fresh transformation (grown on plates for less than 1 week) are used to inoculate 1 ml sterile LBA medium, and these are grown overnight with shaking at 37 °. To 3 ml of fresh sterile LBA is added 60/zl of the overnight culture. This is shaken at 37 ° and monitored spectrophotometricaUy until the OD6o0 = 0.4-0.6 (one should include an extra culture or two from which samples can be removed; these samples lose their sterility during handling and should not be returned to the culture tubes. For simplicity, one can assume that all of the cultures grow at the same rate). At this point, 1 ml of each culture is transferred to a 1.5-ml Eppendorf tube, and in the remaining 2 ml of cells protein synthesis is induced by the addition of 40/xl of 0.1 M aqueous IPTG. The 1-ml uninduced sample is microcentrifuged for 1 min, the supernatant is decanted and discarded, and the cell pellet is frozen at - 2 0 °. The induced cultures are shaken at 37 ° for another 2-6 hr, and then 1 ml of each is transferred to a 1.5-ml Eppendorf tube and microcentrifuged for 1 min. The supernatants are again discarded, and at this point the cell pellets can all be stored at - 20° or directly taken on to the next step. Each cell pellet is resuspended, by repeated pipetting, in 100/xl SDS-PAGE loading buffer. The samples are then heated in a boiling water bath for 3 min (with caps of tubes either left open or weighted down). Samples (1-5 /zl) should be loaded onto an S D S - P A G E gel while still warm to hot, because we have observed that some proteins precipitate on cooling. Samples that are too viscous to be
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
97
pipetted easily should be diluted with an equal volume of SDS-PAGE loading buffer and reheated before loading. A clone transformed with the insert-free overproduction vector should be included in the induction test as a negative control. To facilitate comparison, each uninduced sample is loaded alongside the corresponding induced sample. It is best to avoid overloading the SDS-PAGE gel, because this can sometimes make the induced protein band difficult to identify, especially if the level of induction is modest. We often find it worthwhile to load each sample twice, varying the amount loaded. A minigel apparatus such as the Hoefer MightySmall II is convenient for induction screening, because large numbers of clones can be analyzed rapidly; however, in some cases minigels may have insufficient resolution to permit clear identification of the induced protein band. If one suspects that the recombinant protein may form inclusion bodies in E. coli, it may be worthwhile to screen pellets from induced cells that have been lysed by sonication or French press. In this way, the protein of interest can be significantly enriched and thereby detected more easily (but only if it is insoluble). When using this approach, however, it should be borne in mind that some proteins partition into both soluble and insoluble material. Finally, there is the issue of how many clones to screen. If the plasmid has already been shown by sequencing to contain the desired construct, we typically test four colonies. With transformants obtained directly from ligation of an expression cassette into a vector, we typically test 10-12 clones, even if they have already been analyzed by restriction digestion and shown to contain the desired insert.
Polymerase Chain Reaction-Generated Mutations The error rate of Taq polymerase is a source of serious concern when using any PCR-based method. 7 Although we have observed only one PCRrelated mutation during our numerous overproducer constructions using ECPCR, all such constructs should be sequenced thoroughly. New England BioLabs and Stratagene are now marketing thermophilic DNA polymerase enzymes (Vent 47and Pfu polymerases, respectively) that possesses proofreading (3' ~ 5' exonuclease) activity as well as exceptional thermal stability, which may make these the enzymes of choice for future ECPCR constructions.
47 A. Neuner, H. W. Jannasch, S. Belkin, and K. O. Stetter, Arch. Microbiol. 153, 205 (1990).
98
VECTORS FOR EXPRESSING CLONED GENES
[7l
Troubleshooting
Amplification If no expression-cassette product is visible on an agarose gel, the quality of the enzyme should be checked by running a control PCR reaction (for example, PCR the polylinker of pUC19 using sequencing primers)--we have found the quality of Taq polymerase from different sources to vary greatly. If the enzyme is found to be reasonably active, reaction conditions should be modified as suggested by Innis and Gelfand? 6 Polymerase chain reactions using libraries as donors are more demanding than those using monoclonal donors. If the recommended PCR optimization routines fail when trying to perform PCR from a library, one can either try a different library or try using a nested PCR approach to increase specificity. 7 If side products make isolation or visualization of the desired product difficult, this can sometimes be overcome by repeating the PCR with a sliver of gel containing the desired product. The PCR products should be electrophoresed on a SeaPlaque (FMC) low-melting agarose gel with a minimum of ethidium bromide (1/zl/60 ml agarose solution) and visualized using long-wavelength UV light only. A portion of the gel corresponding to the size of the desired product is excised with a scalpel. The gel slice is then melted in TE buffer and serially diluted to a final value of -10,000fold. The PCR is then repeated, using this diluted DNA as donor. Alternatively, the addition of 1 tzl of I U//zl Perfect Match (Stratagene) can sometimes reduce the amount of undesired products.
Digestion Failure to obtain transformants having the desired insert can result from a number of factors. The most common problem that we have encountered has been the failure of the restriction enzyme to cleave the sites at the ends of expression cassettes. This problem is much worse when the restriction sites are flanked by GC-rich sites such as in a GC clamp; therefore, we strongly recommend the avoidance of such sites in ECPCR. If one suspects a problem in restriction digestion, the best course of action is to end label the cassette with 32p, digest it, and run the products on a 20% (w/v) denaturing polyacrylamide gel. Oligonucleotide size markers corresponding to the sizes of the expected cleavage products should be run in parallel. If this digestion works, scale up all components of the digestion arithmetically, including the enzyme. It is best to bear in mind that restriction digestion of terminal sites is always significantly slower than restriction digestion of internal ones. If the ends are altogether resis-
[7]
OVERPRODUCTION OF PROTEINS USING E C P C R
99
tant to digestion, one should try ligating the PCR product prior to restriction digestion, as described. 48 Another possible source of trouble is that the vector has not been digested adequately. This problem may be difficult to track down, because the polylinker fragment removed during vector preparation is generally difficult to visualize on a gel. For this reason, whenever possible we prefer to digest vectors having a large insert (such as another expression cassette) rather than a polylinker, so that one can readily observe cleavage of the insert on a gel. Ligation Blunt-ended ligation may be effective if one has experienced difficulty preparing the vector or insert for sticky-ended ligation. If the expression cassette is to be used in a blunt-ended ligation, the Klenow fragment of E. coli DNA polymerase (about 0.5 /zl of a 6 U//xl solution) should be added at the end of the PCR (after the reaction has cooled to 37°), and the tube should be incubated at 37 ° for 30 min. This extra step ensures that the Y-A overhangs frequently added by Taq polymerase are removed by the exonucleolytic activity of Klenow. 49~5° All subsequent steps are the same as described above, with the exception that the restriction digestion step is replaced by 5' phosphorylation using T4 polynucleotide kinase. -st Transformation or Loss o f Overproduction High-level overproducers, particularly those present at high copy number, place a severe metabolic tax on the host cell. Inadequate repression of the promoter of the overproducer may lead to problems in cell viability, which may result in failure to obtain overproducing transformants. One should take particular note of this problem when using E. coli promoters that are controlled by the lac repressor (e.g., lac, tac, and trc). Whenever possible, one should transform such overproducers into E. coli cells that overproduce lac repressor, such as lacl Q o r lacI QI strains. As noted above, overproducers should not be maintained on plates, but should be freshly transformed before each growth of new protein. In our hands, plates 48 V. Jung, S. B. Pestka, and S. Pestka, Nucleic Acids Res. 18, 6156 (1990). 49 j. M. Clark, Nucleic Acids Res. 16, 9677 (1988). 50 A. Hemsley, N. Arnheim, M. D. Toney, G. Cortopassi, and D. J. Galas, Nucleic" Acids Res. 17, 6545 (1989). 5t S. Tabor, in "Current Protocols in Molecular Biology" (F. M. Ausubel, R. Brent, R. E, Kingston, D. D. Moore, J. A. Smith, J. G. Seidman, and K. Struhl, eds.), Vol. 1, Unit 3.10, p. 2. Wiley, New York, 1987.
100
VECTORS FOR EXPRESSING CLONED GENES
[7]
freshly streaked from glycerol stocks are not as reliable as freshly transformed cells. Poor Levels of Expression Not every gene can be overexpressed at high levels in E. coli. In particular, genes having codons that are used at low frequency in E. coli may be inherently difficult to express. In many other cases, however, expression can be increased by changing the RBS. We recommend that one try one or all of the RBS sequences in Fig. 3A. Because ECPCR allows one to carry out several overproducer constructions in parallel, we sometimes screen all of the RBS sequences in Fig. 3A, and use induction testing to determine the efficiency of each. Insoluble Protein On occasion, proteins overexpressed in E. coli fail to undergo proper folding to yield soluble, active protein. This can be readily identified on lysis of induced cells followed by centrifugation, whereupon the desired protein is found in the pellet fraction rather than in the supernatant. However, before concluding that the protein is insoluble, one should first try to wash the lysis pellet with 1 M salt, which will release proteins that are active but tightly bound to chromosomal DNA (as with many DNAbinding proteins). The various strategies for recovery of active folded protein from these insoluble aggregates have been reviewed by Marston. 52-54
Concluding Remarks ECPCR offers a rapid and reliable means for the construction of E. coli overproducers. This method requires little specialized training in molecular biology, such that we have found it to be rapidly mastered even by undergraduate students with no prior experience. By taking advantage of the ability of PCR to modify DNA sequences rapidly, ECPCR considerably reduces the time required to construct an overproducer: we have frequently completed the entire process from primer synthesis to overproducer sequencing in a week and, in our hands, the process has almost always been completed within 2 weeks. Another advantage of ECPCR is 52 F. A. O. Marston, Biochem. J. 240, 1 (1986). 53 F. A. O. Marston, in " D N A Cloning" (D. M. Glover, ed.), Vol. 3, p. 59. IRL Press, Washington, D.C., 1987. 54 C. H. Schein, Biotechnology 7, 1141 (1989).
[71
OVERPRODUCTION OF PROTEINS USING E C P C R
101
that donor DNA present in small quantities or in a library may be used. This feature has the significant benefit of eliminating the need to physically transfer clones from one laboratory to another. The generality of ECPCR is amply demonstrated by the broad range of proteins that have been overproduced using it: inter alia human CD4 2 and intracellular adhesion molecule (ICAM)-155 domains, human cyclophilins A 28 and B :6 E. coli cyclophilin, 57 human FK506 binding protein (FKBP- 12)58 and FKBP- 13,59 E. coli histidine-tRNA synthetase,6° human arylamine N-acetyltransferases NAT1 and NAT2, 61 E. coli D-alanine ligases Ddla and Ddlb, 6z Trypanosoma cruzi trypanothione reductase, 63 human transcription factor KBF1, 64 E. coli DNA repair proteins Ada, AlkA, and Fpg, 29Haernophilus aegyptus DNA cytosine 5-methyltransferase M.HaelII, 65 E. coli CMP-NeuAc synthetase, 66 Bacillus subtilis DNA polymerase 1II, 67 and the homeodomain of Saccharomyces cerevisiae a2 repressor. 68 In principle ECPCR should allow the overexpression of proteins in other organisms as well. Of particular interest in this regard are the transcription vectors that have been developed for expression in baculovirus.69 Although not much is yet known about optimal sequences for translation initiation in baculovirus, as more information becomes available we expect ECPCR will provide a powerful means by which to maximize expression in this eukaryotic system. We and others are exploring useful variations on ECPCR. In one, which we term selectable ECPCR, degenerate RBS sequences are used in s5 K. D. MacFerrin, S. L. Schreiber, and G. L. Verdine, unpublished observations. 56 E. R, Price, L. D. Zydowski, M. Jin, C. H. Baker, F. D. McKeon, and C. T. Walsh, Proc. Natl, Acad. Sci. U.S.A. 88, 1903 (1991). 57 j. Liu and C. T. Walsh, Proc, Natl. Acad. Sci. U.S.A. 87, 4028 (1990). 5s R. F. Standaert, A. Galat, G. L. Verdine, and S. L. Schreiber, Nature (London) 346, 671 (1990). 59 p. Martin and S. L. Schreiber, unpublished observations. C. Francklyn and P. Schimmel, Proc. Natl. Acad. Sci. U.S.A. 87, 8655 (1990). 61 D. M. Grant and J. M. Dupret, FASEB J. 5, 1205 (1991). 6z L. E. Zawadske, T. D. H. Bugg, and C. T. Walsh, Biochemistry 30, 1673 (1991). 63 F. X. Sullivan and C. T. Walsh, Mol. Biochem. Parasitol. 44, 145 (1991). M. Sodeoka, C. J. Larson, L. Chen, K. P. LeClair, W. S. Lane, and G. L. Verdine. submitted for publication. 65 L. Chen, A. M. MacMillan, W. Chang, K. Ezaz-Nikpay, W. S. Lane, and G. L. Verdine, Biochemistry 30, 11081 (1991). 66 y. [chikawa, G.-J. Shen, and C.-H. Wong, J. Am. Chem. Soc. 113, 4698 (1991). 67 R. A. Hammond, M, H. Barnes, S. L. Mack, J. A. Mitchener, and N. C. Brown, Gene 98, 29 (1991). 68 C. L. Phillips, A. K. Vershon, A. D. Johnson, and F. W. Dahlquist, Genes Dev. 5, 764
(1991). 69 N. R. Webb and M. D. Summers, Technique 2, 173 (1990).
102
VECTORS FOR EXPRESSING CLONED GENES
[8l
the 5' primer, and the resulting library of transformants is screened or selected for those that o v e r p r o d u c e at the highest level. 29 In another, which we h a v e termed leapfrog E C P C R , nested primers are used to add long stretches of sequence information to the donor sequence during PCR. 55 Particularly valuable extensions of E C P C R involve the addition of periplasmic export sequences, 55 cysteine " h a n d l e s " for specific chemical modification, 29 or peptide " t a g s " for affinity purification to proteins. 66 Finally, we wish to note that methods similar to E C P C R have b e e n d e v e l o p e d for use in the in vitro production of proteins. While the amounts of protein obtained are generally small, these methods require no cloning. 70-7~
Acknowledgments We wish to thank members of the Schreiber and Verdine groups, especially H. Nash, R. Standaert, J. Liu, K, Ezaz-Nikpay, and P. Martin for helpful discussions and comments. We are grateful to T. Ellenberger and the Harrison group for sharing their protocol for purification of synthetic oligonucleotides. We thank S. Burakoff, B. Bierer, S. Reddy, B. Cochran, B. Seed, M. L. Hedley, T. Maniatis, M. Ptashne, and M. Sekiguchi for supplying materials. R. Standaen provided expert experimental assistance. We thank Professor D. E. Cane and C.-H. Wong for sharing results prior to publication. K.D.M. is the recipient of a National Science Foundation graduate fellowship. This work was supported by the National Institutes of Health (Grant GM30738-09, to S.L.S.). and by fellowships from the NSF (Presidential Young Investigator Program), the Searle Scholar Program, Alfred P. Sloan Foundation, Eli Lilly, Hoffmann-La Roche, and Bristol Myers-Squibb (to G.L.V.). 7oC. Wychowski, S. U. Emerson, J. Silver, and S. M. Feinstone, Nucleic Acids Res. 18, 913 (1990). 71E. R. Mackow, M. Y. Yamanaka, M. N. Dang, and H. B. Greenberg, Proc. Natl. Acad. Sci. U.S.A. 87, 518 (1990). 72 K. C. Kain, P. A. Orlandi, and D. E. Lanar, BioTechniques 10, 366 (1991).
[8] E f f i c i e n t C o m p l e m e n t a r y Expression
Using Polymerase
DNA
Amplification
Chain Reaction
and
Technology
B y THOMAS L. PAULS and MARTIN W. BERCHTOLD
Introduction and Principles of Methods E x p r e s s i o n of foreign genes in bacteria is a powerful method in molecular biology that allows the a c h i e v e m e n t of a variety of goals, including the production of large a m o u n t s o f rare proteins or peptides to be used for basic research or directly for therapeutic applications. In addition, once a
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
[8]
cDNA CLONINGBY PCR
103
protein can be expressed in bacteria, site-directed mutagenesis is often carried out to establish structure-function relationships. Many different protocols are available to generate specific cDNA and clone these molecules into appropriate expression vectors. In this chapter we describe a rapid and efficient method that takes advantage of the speed and efficiency of polymerase chain reaction (PCR) technology for the steps mentioned above. Since PCR was first proposed and shown to be useful as a routine laboratory tool in 19851 this technique has been rapidly introduced in research and diagnostic laboratories in academia, industry, and hospitals. Literature dealing with technical improvements and novel practical applications as well as simply with the utilization of standard PCR procedure is increasing dramatically (for collections of applications and protocols, see Refs. 2-4). The PCR is not only the method of choice for directly amplifying minute quantities of genomic DNA from a complex background but has also been used successfully to amplify specific cDNA molecules obtained by reverse transcription of complex mixtures of mRNA. 4 So far, application of the PCR to isolate specific cDNA has required knowledge of DNA sequences flanking the region of interest. Our laboratory 5'6 and others 7'8 have proposed PCR methods that use only one specific oligonucleotide and a second unspecific primer. This is particularly interesting if 3' and 5' flanking regions of a coding sequence of interest are not known. Using this method (outlined in Fig. 1) we have cloned the 3' part of human parvalbumin cDNA 9 and the full-length cDNA for the human ribosomal large subunit protein L17A. 1° Subcloning of PCR products is often a problem. Restriction sites at the flanking sequences of PCR products introduced by attachment at the 5' end of the oligonucleotides are in many cases efficiently cleaved only when several additional nucleotides I R. K. Saiki, S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. Arnheim, Science 230, 1350 (1985). ' H. A. Erlich (ed.), "PCR Technology." Stockton, New York, 1989. 3 M. A. lnnis, D. H. Gelfand, J. J. Sninsky, and T. J. White eds., "PCR Protocols" Academic Press, Inc., San Diego (1990). 4 E. S. Kawasaki and A. M. Wang, in "PCR Technology" (H. A. Erlich, ed.), pp. 89-97. Stockton, New York, 1989. s M. W, Berchtold, Nucleic Acids Res. 17, 453 (1989). T. M. J. Leu, C. C. Kuenzle, and M. W. Berchtold, in "Methods in Gene Technology" (J. W. Dale and P. G. Sanders, eds.), pp. 239-265. JAI Press, London, 1990. 7 M. A. Frohman, M. K. Dush, and G. R. Martin, Proc. Natl. Acad. Sci. U.S.A. 85, 8998 (1988). O. Ohara, R. L. Dorit, and W. Gilbert, Proc. Natl. Acad. Sci. U.S.A, 86, 5673 (1989). 9 M. Berchtold, J. Mol. Biol. 201, 417 (1989). l0 M. Berchtold and M. C. Berger, Gene 102, 283 (1991).
104
VECTORS FOR EXPRESSING CLONED GENES
5'
3'
(A) n 3'
re,erse
5' directed PCR
transcription]
I
1
3' directed PCR ,
(T)25 5' I
I
[8]
(T)25
5'
tailing by [ [terminal tranaferaae[ 2
3' (A)50_100 1
[
1
(T)25 5' 3
PCR 20 cyc es
1 4 "nested" PCR 25 cycea [
30-40 cyclesl [PCR
÷
t
__\
¢
I o'on'o0'oto" direct sequencing
J__ [
FIG. 1. General strategy for 3'- and 5'-directed PCR. I f only minimal information on a sequence of interest [down to 20 nucleotides (nt)], which may be mixed owing to codon degeneracy, is available, full-length cloning can be achieved by the following strategy. After reverse transcription using oligo(dT) as primer, 3'-directed PCR with a gene-specific primer and oligo(dT) is carried out (shown on the right). The new sequence information obtained from sequencing the PCR product is used to design primers for 5'-directed PCR (shown on the left). The template for this reaction is dA-tailed eDNA (by terminal transferase). Two rounds of Y-directed PCR with a °'nested" primer for the second round are generally needed to obtain specific PCR products that can be sequenced directly or after cloning into the desired vectors. Horizontal arrows indicate the direction of DNA synthesis. Primer 1, oligo(dT) with additional sequence; primer 2, gene-specific primer for 3'-directed PCR; primer 3, gene-specific primer for 5'-directed PCR; primer 4, " n e s t e d " primer for Y-directed PCR. Open squares, additional " a n c h o r e d " primer sequences (25 nt); blackened rectangles, XbaI restriction site; arrowhead at the 5' ends of the primers and in the PCR products, SacI restriction site.
[8]
cDNA CLONINGBY PCR
105
(e.g., GC clamps) are added, ll Several other possibilities exist to avoid the problem of cloning PCR products. In the TA cloning procedure, dA that is normally added to the PCR product by the intrinsic terminal transferase activity of Taq polymerase 12 is annealed to a dT tail at both ends of the vector that have been artificially added by terminal transferase 13 or by the Taq polymerase.14 A possibility for increasing the restriction enzyme efficiency on both ends of the PCR product is to create concatamers of the PCR product by self-ligation followed by digestion with restriction enzymes.15 A ligase-free subcloning method has been published.16 Sequences preceding the ATG initiation sites and that are known to allow optimal translation can be incorporated into the PCR product by including them at the 5' end of the oligonucleotide. For our expression system we took advantage of the naturally highly efficient initiation site of T7 bacteriophage gene 10 that encodes the T7 major capsid protein and is present in the pGEMEX vector. For precise insertion of the coding sequence of interest into this vector, PCR is carried out with a 5' primer containing the ATG codon within an N d e I site that is also present in the vector (Fig. 2). The final part of this chapter deals with the heterologous expression of genes in bacteria under the control of the inducible T7 promoter in a modified pGEMEX vector according to Studier et al. J7 and quantitation as well as isolation of the recombinant protein from crude bacterial extracts. Procedures are described in the order that they are carried out in a typical experiment, starting with a tissue or a cell culture. As an experimental test system we chose to amplify and clone the cDNA for the rat Ca2+-binding protein parvalbumin, 18using rat brain as a source for RNA preparation. Conditions for protein expression, quantitation, and isolation were used to achieve parvalbumin production with high efficiency. For detailed information concerning reagents and preparation of solutions see Materials and Reagents (below). For standard molecular biology procedures, consult the laboratory manual by Sambrook et al.19 tt D. L. Kaufman and G. A. Evans, BioTechniques 9, 305 (1990). t~ j. M. Clark, Nucleic Acids Res. 16, 9677 (1988). i_~T. A. Holton and M. W. Graham, Nucleic Acids Res. 19, 1156 (1991). 14 D. Marchuk, M. Drumm, A. Saulino, and F. S. Collins, Nucleic Acids Res. 19, 1154 (1991). 15 V. ]ung, S. B. Pestka, and S. Pestka, Nucleic Acids Res. 18, 6156 (1990). 16 A. R. Shuldiner, L. A. Scott, and J. Ross, Nucleic Acids Res. 18, 1920 (1990). 17 F. W. Studier, A. H. Rosenberg, J. J. Dunn, and J. W. Dubendorff, this series, Vol. 185, p. 60. ~s M. W. Berchtold, Biochim. Biophys. Acta 1009, 201 (1989). ~9j. Sambrook, E. F. Fritsch, and T. Maniatis, "'Molecular Cloning: A Laboratory Manual." 2rid Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989.
106
VECTORS FOR EXPRESSING CLONED GENES ATG
[8]
TAA
I
I mRNA
I 0.06
"I'
~
__Sac I 0.07
T7 terminator
~
~-
NdQI
PCR
MEX paEMEx 4.00~-k,-~ Kb
gene 10 \~L__..Nde i 0.91 ',q=_ / T7 promoter
P ORJ ATG
TAA
Sac~
Nd~I
Ligation
FIG. 2. Strategy for cloning of PCR products into expression vectors. Rat parvalbumin cDNA was amplified by PCR after reverse transcription of brain mRNA. Two gene-specific primers (bold arrows) containing the indicated restriction sites were used. After cutting the PCR product and the vector with NdeI and SacI restriction enzymes, ligation of these molecules was carried out to obtain the expression vector pGEMEX-PV.
[8]
cDNA CLOYINGBY PCR
107
Material and Reagents For R N A Isolation
Guanidinium denaturing solution: 4 M guanidinium isothiocyanate (Fluka, Ronkonkoma, NY), 25 mM sodium citrate, pH 7.0 (adjusted with acetic acid), 0.5% (w/v) N-laurylsarcosine. Before use, add 150 tzl 2-mercaptoethanol to 20 ml of the above mixture Sodium acetate, 3 M, pH 4.0 (adjusted with acetic acid): Suspend sodium acetate in a minimum of water and dissolve by adding glacial acetic acid until the desired pH is attained. Adjust volume with water Phenol for RNA extraction: Mix liquid phenol with 1 vol of water. Upper (aqueous) layer should have pH 4.0. Store at 4° (do not use for DNA treatment) Chloroform-isoamyl alcohol mixture (49: 1, v/v) TE buffer, pH 7.5:10 mM Tris-HCl, pH 7.5, 1 mM EDTA 2-Propanol LiC1, 4 M Sodium dodecyl sulfate (SDS), 10% (w/v) For Reverse Transcription
Reverse transcriptase: Cloned, from Moloney murine leukemia virus (Mo-MuLV) (13 U/tA) (Bethesda Research Laboratories, Gaithersburg, MD) Reverse transcription reaction buffer [5 × concentrated: 250 mM TrisHCI, pH 8.3 (at 40°), 150 mM KCI, 40 mM MgC12,50 mM dithiothreitol (DTT)], stored at - 2 0 ° Primer for reverse transcription: Oligo(dT) with additional primer 5'-GCAGCTTGCGACGGTGTCTAGACGC(T)25 or defined genespecific primers (10 p~Meach). Primers are generally purified on a 15% (w/v) polyacrylamide gel containing 7 M urea, followed by reversedphase chromatography using a SepPak CI8 cartridge (Waters Chromatography Division, Millipore Corporation, Milford, MA), lyophilization, and redissolving in HzO dNTP stock solution: dATP, dCTP, dGTP, dTTP (ultrapure; Pharmacia, Piscataway, NJ) in one stock solution; 10 mM each in TE buffer Ribonuclease inhibitor (RNasin), 50 U/~I (Boehringer Mannheim Biochemicals, Indianapolis, IN) Hydrolyzing solution: 0.3 N NaOH, 30 mM ethylenediaminetetraacetic acid (EDTA)
108
VECTORS FOR EXPRESSING CLONED GENES
[8]
Organic solvents: Phenol for DNA extraction (preparation as described in Ref. 19), dichloromethane, 100% ethanol, 2-propanol DNA terminal transferase kit (Boehringer Mannheim)
For Polymerase Chain Reaction Polymerase chain reaction buffer, 5 × concentrated (5 × PCRB): 83 mM ammonium sulfate, 335 mM Tris-HC1, pH 8.8 at 25°, 50 mM 2-mercaptoethanol, 1 mg/ml bovine serum albumin (BSA) (Dnasefree quality; Boehringer Mannheim) dNTP stock solution: As described for reverse transcription, above, but at a concentration of 1 mM each AmpliTaq (cloned Taq DNA polymerase; Perkin-Elmer Cetus, Norwalk, CT), 5 U//xI: Prior to use, dilutions of 0.2 U//zl are made in I x PCRB TEC buffer: Carrier DNA poly(dI-dC), poly(dI-dC), 25 /zg/tzl (Pharmacia) in TE, pH 7.5 Mineral oil, extra light (Sigma, St. Louis, MO): Autoclaving is not necessary Primers (purified as described above, 10 tzM stocks) according to the type of PCR to be carried out
For Cloning of Polymerase Chain Reaction Products Glycogen (20 mg/ml), DNAse free (Boehringer Mannheim) TE buffer, pH 8.0:10 mM Tris-HCl, pH 8.0, 1 mM EDTA NdeI (I0 U/p.l) and SacI (12.5 U//zl) restriction enzymes (Boehringer Mannheim) Organic solvents: Phenol for DNA extraction (preparation as described in Ref. 19), 100% ethanol Agarose and other material for gel electrophoresis (as described in Ref. 19) pGEMEX vector, obtained from Promega (Madison, WI), modified by M. Ziak and R. Jaussi to eliminate a second NdeI site at position 3259 ATP stock solution, 10 mM (Pharmacia) DTT, 100 mM (Bio-Rad, Richmond, CA) T4 ligase, 5 U//zl (Pharmacia) Ligation buffer, 10 x : 20 mM Tris-HCl, pH 7.6, 50 mM MgC12 Spermidine, 10 mM (Sigma)
[8]
cDNA CLONINGBY PCR
109
Reagents for transfection (see Ref. 20) Reagents for sequencing (see Ref. 19) Generation of eDNA The starting point is a tissue or a cell culture that contains a mRNA encoding a protein of interest to be expressed in bacteria. Only limited sequence information of this mRNA is required. Several criteria are important for the evaluation of an optimal RNA isolation procedure. RNA must be protected against degradation by RNase present in tissues and on laboratory material. The yield of purified RNA should be high and reproducible, and the method should be fast and easy. Several companies offer ready-to-use kits for the isolation of total and poly(A) + RNA. We generally obtain good results using a single-step extraction with guanidinium thiocyanate-acidic phenol-chloroform, 21 followed by lithium chloride precipitation. 22
Experimental Procedure RNA Isolation. Fresh or frozen ( - 8 0 °) tissues or cultured cells (100 mg wet weight) are homogenized at room temperature in 1 ml guanidinium denaturing solution using a glass or Teflon homogenizer. To the homogenate, 100 ~1 of a chloroform-isoamyl alcohol mixture is added, followed by vortexing for at least 1 min. This solution is centrifuged at 10,000 g for 20 min at 4 °. The upper phase containing RNA is saved and 1 ml phenol (pH 4.0) and 200/~1 chloroform-isoamyl alcohol mixture are added. This solution is vortexed, followed by centrifugation at 10,000 g for 20 min at 4°. RNA is precipitated from the upper phase by addition of an equal volume of 2-propanol ( - 20°, 1 hr). RNA is collected by centrifugation at 15,000 g for 20 min at 4 ° and redissolved in 300 ~1 guanidinium denaturing solution followed by reprecipitation by adding 2-propanol as described above. The RNA pellet is resuspended in 200 /A of 4 M LiCl to solubilize polysaccharides. The insoluble RNA is pelleted at 10,000 g for 10 min at 4°. This pellet is dissolved in 200 p,1 of TE buffer, followed by addition of 10 IA SDS (10%, w/v) and 200/A chloroform-isoamyl alcohol mixture and centrifugation at 3000 g for 10 rain at 4 °. The aqueous phase is precipitated with 200 ~1 2-propanol and 200 p.l 3 M sodium acetate, pH 4.0. The RNA pellet is dissolved in 50/.d H20 and the optical density at h260 and h280 is 20 D. Hanahan, J. Mol. Biol. 166, 557 (1983). zE p. Chomczynski and N. Sacci, Anal. Biochem. 162, 156 (1987). -'2 C. Puissant and L.-M. Houdebine, BioTechniques 8, 148 (1990).
1 I0
VECTORS FOR EXPRESSING CLONED GENES
[8]
measured. Good-quality RNA has an OD260/280 of 1.6 to 1.9 ( 1 0 D at h260 equals 40 ~g RNA/ml). We generally obtain yields in the range of 150-200 /zg RNA/100 mg of tissue. Notes: To avoid RNase contamination, the use of latex gloves is recommended. Solutions and laboratory equipment should be sterilized, autoclaved, or treated with 0.1 N NaOH followed by rinsing with distilled water. Generally, total or cellular RNA can be directly used for reverse transcription without poly(A) + enrichment. If this is desired, for example, if the mRNA of interest is present in low abundance for special applications, standard oligo(dT) affinity purification protocols as described in Ref. 19 may be applied. Reverse Transcription Reaction. In a final volume of 25/.d the following components are assembled: Reverse transcription reaction buffer (5 x concentrated), 5/.d dNTP stock solution (10 mM each), 2.5/zl RNasin (50 U/ml), 0.25/zl Primer (10/zM), 1 /zl RNA [total cytoplasmic or poly(A)+], 0.1-5 tzg pretreated for 3 min at 70 ° and quenched on ice, 10/zl Reverse transcriptase (Mo-MuLV, cloned) (13 U//zl), 2.5/zl Distilled H20, 3.75/zl The reaction is carried out at 42 ° for 60 min and stopped by adding 25 tzl of hydrolyzing solution. The sample is neutralized by adding 12.5/zl of 1 M Tris-HCl, pH 8.0. The cDNA is phenol extracted by adding 65/xl phenol-dichloromethane mixture (I : 1, v/v) once and the upper phase is washed with 65/xl dichloromethane followed by centrifugation. The cDNA in the upper phase is precipitated after adding 1 M NaCI to a concentration of 100 mM, followed by mixing with 130/~1 ethanol. After cooling at - 2 0 ° for 1 hr the cDNA is collected by centrifugation at 15,000 g for 20 min at 4°. The pellet is washed with 75% (v/v) ethanol and dried, followed by redissolving in TE buffer to a concentration of 1-25/zg cDNA/Izl. Notes: It may be advisable to reduce the primer concentration by a factor of 102 to 103 if a gene-specific primer is used for the reverse transcription of a low-abundance transcript to avoid unspecific priming. For calculation of dNTP incorporation, one of the dNTPs is added in the 3Zp-labeled form as described in Ref. 19. It is not necessary to remove the RNA after reverse transcription if PCR is followed immediately. If cDNA must be stored we recommend alkali treatment as described above. If small amounts of RNA (<0.5/.~g) are used it may be necessary to
[8]
cDNA CLONINGBY PCR
111
precipitate the produced cDNA by the use of 5/zg tRNA (phenol extracted according to Ref. 19). Polymerase Chain Reaction. Analytical PCR is carried out in 10/xl and preparative PCR in 100/xl PCR buffer. A "typical" analytical PCR contains the following components: PCRB (5 x ), 2/zl dNTP stock solution (1 mM each), 2/xl MgCI 2 (15 mM), 1 /zl Primer 1 (10/~M), 1/xl Primer 2 (I0 IxM), 1 /zl Taq polymerase (AmpliTaq, 0.2 U/tzl), 1 tzl Template (0.1-1 ~g cDNA//A TEC buffer), 1 /zl Distilled H20, 1 /zl The samples are covered with a drop of mineral oil. Depending on the abundance of the specific cDNA to be amplified, 20 to 40 cycles are performed. Excessive cycle numbers often lead to a decrease in yield and efficiency. As a standard protocol to generate products smaller than 1000 bp the following conditions are applied: 95 ° melting for 30 sec, 50° annealing for 30 sec, 73 ° synthesis for 1 min. These conditions must be optimized, depending on the properties of the PCR cycler, the length and GC content of the primers, and the expected product length. Three types of PCR (internal, 3' directed, and 5' directed) are described as outlined in Fig. 1. Internal polymerase chain reaction: Two gene-specific primers are used for this classical type of PCR for a single round of PCR. 3'-Directed polymerase chain reaction: The reaction mixture contains, along with the usual components of the PCR mix, a single-stranded cDNA template (in 1/xl TEC buffer) carrying a (dT)25 extension on its 5' end and a gene-specific PCR primer complementary to the template sequence with a restriction site and the nonspecific primer, (dT)25. Typically, PCR is carried out over 30-40 cycles. If no specific PCR products are obtained, a second round of PCR with a second, "nested" gene-specific primer that is placed 3' of the first, gene-specific primer may be carried out. 5'-Directed polymerase chain reaction: For 5'-directed PCR the cDNA must be tailed at its 3' end. We generally use dA for tailing. Conditions should be applied that produce a poly(dA) tail of about 50-100 nucleotides. Ten to 50 ng of single-stranded cDNA is used for the reaction. The Boehringer Mannheim terminal transferase kit contains all the necessary reagents. Conditions indicated in the manual should be applied. Two rounds of PCR are generally required for optimal specificity. In 23 A. P. Feinberg and B. Vogelstein, Anal. Biochem. 132, 6 (1983).
112
VECTORS FOR EXPRESSING CLONED GENES
[8]
the first round, the single-stranded cDNA template (in 1/xl TEC buffer) is a reverse transcript of an mRNA, which is generated by the use of either oligo(dT) or a gene-specific oligonucleotide primer, and which carries a poly(dA) extension on its 3' end, generated by terminal transferase. This template is added to 9/zl of a premix that contains, along with the usual components, a first gene-specific primer (identical to the cDNA sequence) and the nonspecific primer, (dT)zs. Typically, this first round of PCR is carried out over 20 cycles. For the second round of PCR, the products of the first round are diluted 10- to 100-fold with H20, and 1 /zl is transferred to a fresh reaction tube (preloaded with 9/zl of premix) to serve as a new template. This premix contains, along with the usual components, a second, nested gene-specific PCR primer and the nonspecific primer, (dT)25. Typically, the second round is carried out over 25 cycles. Notes: DNA carryover is a major problem in all types of PCR. To reduce carryover contamination it is recommended that the reagents be divided into aliquots to be used only once or twice, that pre- and postPCR operations be separated physically, and that different pipettes be used for pre- and post-PCR work. Ultraviolet (UV) irradiation of PCR reagents (except DNA) may be helpful to prevent DNA carryover. A control PCR in the absence of template is recommended to check for contaminated solutions. In our hands, the use of oligo(dT) with an additional 5' extension of 25 randomly selected nucleotides for reverse transcription 3'- and 5'-directed PCR gives superior product specificity. The sequence successfully used for our experiments is 5'-GCAGCTTGCGACGGTGTCTAGACGC(T)25, carrying an XbaI site at its 5' end. For convenience this oligonucleotide is referred to as oligo(dT) in text. However, when only the 25-nucleotide (nt) 5' extension was used as a primer starting with the second round of the PCR cycle no further increase in specificity was observed. MgCI 2 and dNTP concentrations are critical for obtaining the desired PCR products. These two parameters should be varied (with final concentrations of MgCI2,1-10 mM and dNTP, 50-250/zM) if yield and specificity are not sufficient. In addition, optimizing the annealing temperature may increase PCR efficiency. It is advisable to run control PCRs with a single primer to check specificity. For PCR, different heat-stable polymerases were tested: Taq polymerase AmpliTaq (Perkin-Elmer Cetus), Taq polymerase (Boehringer Mannheim), and Replitherm (Epicentre Technologies, Madison, WI). All gave basically the same results. These polymerases have different Mg 2÷ requirements. Analysis of polyrnerase chain reaction products: The reaction mixtures can be applied directly to polyacrylamide or agarose gels depending on the expected product length. DNA is stained with ethidium bromide and
[8]
cDNA CLONINGBY PCR
1l 3
visualized at 300-nm UV light according to standard procedures.~9 The specificity of PCR products may be tested by Southern blot analysis.~9 Transfer of DNA from polyacrylamide gels (5%, w/v) to nylon membranes is conveniently done by alkaline DNA blotting. To probe for specific products of 3'- and 5'-directed PCR, radiolabeled products of an internal PCR can be used, provided the internal product extends considerably into the sequence delineated by the nested primer used in the second round of PCR. To obtain DNA probes of high specific activities, the random oligonucleotide-labeling method 23 is used. It is highly recommended that PCR products be sequenced before continuing further work. Many different methods exist for sequencing PCR products either directly or after cloning into an appropriate vector (reviewed in Ref. 24). Cloning of Polymerase Chain Reaction Products. The required steps for cloning of PCR products into an expression vector are outlined in Fig. 2. As an example, cloning of the rat parvalbumin cDNA into pGEMEX is shown, mRNA from brain tissue is used as a template for reverse transcription. Preparative PCR (100/xl in two 50-/xl portions) is carried out under standard conditions with the following primers: The upstream primer is 5'-CCCGGATCCAAGTTGCCATATGTCGATG, corresponding to rat parvalbumin cDNA positions 56-8225 with a BamHI site near the 5' end and an NdeI site in the center. The downstream primer is 5'GAGGTCGACGAGCTCGGTACCAAGCAGGCAGGA with a SacI site near its 5' end. The cycler was programmed as follows: 95 ° melting for 40 sec, 50° annealing for 40 sec, 73 ° synthesis for 1 min, 30 cycles. The PCR product is run on an 8% (w/v) polyacrylamide gel. The gel piece with the PCR product of the correct size (689 bp) is then dissected and crushed into small pieces using a Teflon rod. The DNA is eluted with 8 voi of H20 overnight at room temperature on a shaker. Gel fragments are separated by centrifugation for 5 min at high speed, using an Eppendorf microfuge. The pellet is resuspended in 2 vol of H20 and centrifuged as above. Both supernatants are pooled. The DNA is precipitated, using glycogen as a carrier (20 tzg/ml). The DNA is resuspended in TE buffer, pH 8.0 and is digested with NdeI and SacI restriction enzymes. For a detailed description of restriction enzyme digestion, phenol extraction, and ethanol precipitation see Sambrook et al. 19 After digestion and phenol extraction the DNA is ethanol precipitated and resuspended in TE. The NdeI/SacIdigested PCR product is purified by preparative polyacrylamide gel electrophoresis as described above. An aliquot of the purified PCR product is 24 U. B. Gyllenstein, BioTechniques 7, 700 (1989). 25 M. W. Berchtold, P. Epstein, A. L. Beaudet, M. E. Payne, C. W. Heizmann, and A. R. Means, J. Biol. Chem. 262, 8696 (1987).
114
VECTORS FOR EXPRESSING CLONED GENES
[8]
run on a 1-2% (w/v) agarose gel for estimation of the DNA concentration. The pGEMEX vector is prepared as follows: Initially, the vector is digested with NdeI and SacI restriction enzymes. Separation of the pGEMEX vector (3200 bp) from the T7 gene 10 sequence (800 bp) is carried out by 1% (w/v) agarose gel electrophoresis and subsequent GeneClean (B10 101, La Jolla, CA) purification of the 3200-bp fragment. Ligation and transfection: For ligation the NdeI/SacI-digested, truncated pGEMEX vector and the PCR product are mixed with different molar ratios of I : 1, 1 : 2, and 2 : 1 (80 ng : 20 ng, 160 ng: 20 ng, 40 ng: 20 ng). A typical ligation mixture contains the following: Vector/PCR product mix, 8.5 tA Ligation buffer (10x), 1.5 IA Spermidine (10 mM), 1.5/zL ATP (10 mM), 1.5/zl DTT (100 mM), 1.5/zl T4 DNA ligase (5 U//zl), 0.5/xl The ligation mix is incubated at 15° overnight. Ligation mix (7/zl) is used for transfection into Escherichia coli JM 101 according to Hanahan. 2° Colonies are characterized by restriction enzyme mapping. Clones with inserted PCR products are sequenced using the dideoxy-sequencing method according to Ref. 19. Notes: Primers should exceed the restriction site by several bases to ensure effective digestion. The PCR products are purified by polyacrylamide gel electrophoresis because of inhibitory effects of traces of PCR reagents on the NdeI restriction enzyme. Isolation of PCR products by agarose gel electrophoresis and GeneClean causes substantial inhibition of NdeI digestion. Expression of Cloned Polymerase Chain Reaction Products, Protein Quantitation, and Isolation The pGEMEX expression vector from Promega is based on the T7 bacteriophage polymerase/promoter expression system developed by Studier et al. 17 High-level expression of cloned genes in vivo is obtained by an optimized vector/host combination. One major advantage of this system is the stringent specificity of T7 RNA polymerase for its own promoter, which is not utilized by E. coli RNA polymerase. T7 RNA polymerase initiates transcription efficiently and elongation is about fivefold faster in comparison to E. coli RNA polymerase. Furthermore, T7 RNA polymerase is relatively unaffected by the presence of non-T7 transcription terminators and thus is able to produce complete transcripts from almost any DNA linked to a T7 promoter.
[8]
cDNA CLONINGBY PCR
115
p G E M E X Vector This vector was developed by Promega for expressing cloned DNA under the control of the T7 promoter from bacteriophage T7 (Fig. 2). This promoter, along with the ribosome-binding site ofT7 gene I0, is positioned upstream from a T7 gene 10 leader sequence encoding the amino-terminal part of the major capsid protein. A T7 transcription terminator is incorporated downstream from the multiple cloning site in the original cloning procedures. Sequences cloned into pGEMEX vector polylinker sites in the correct reading frame are expressed as T7 gene 10 fusion proteins with a gene 10 product amino-terminal part of 260 amino acids. Alternatively, the gene of interest can also be expressed without a gene 10 fusion part. Taking advantage of an NdeI restriction site at the initiation codon of gene 10, the T7 gene 10 fragment can be directly replaced by any DNA of interest, as long as the new DNA fragment contains a NdeI restriction site (CATATG) at the ATG initiation codon and any restriction site at the 3' end that is compatible with the pGEMEX polylinker. A novel NdeI restriction site, if not present in the DNA, can be introduced conveniently by site-directed mutagenesis using conventional methods or PCR. pGEMEX contains, in addition to the NdeI site at the gene 10 translation initiation site, a second NdeI restriction site at position 3259. The latter site was eliminated by partial digestion of the plasmid and religation after creation of blunt ends (M. Ziak and R. Jaussi, University of Ztirich, Switzerland, personal communication 1991). pGEMEX contains an ampicillin resistance gene for selection of transfected bacteria and the fl origin of replication in order to generate singlestranded DNA for sequencing and mutagenesis. Host Strains for Expression The bacterial hosts for expression are E. coli JM 109 (DE3) [recAl supE44 endAl hsdR17 gyrA96 relA1 thi A(lac/prbAB) F'(traD36 proAB ÷ lacl 9 lacI 9 lacZ AM15)] from Promega and BL 21 (DE3) pLysE [hsds gal (~cIts857 indl Sam7 nin5 lacUV5-T7 gene 1)], kindly provided by Dr. W. Studier (Brookhaven, Upton, NY). Both strains were constructed by cloning T7 gene 1 (encoding T7 RNA polymerase) into the ~ vector D 69 and then integrating this DNA into the chromosome of the two E. coli strains.IV The T7 RNA polymerase gene is under control of the lacUV5 promoter and is therefore inducible by addition of isopropyl-/3-D-thiogalactopyranoside (IPTG) to the culture medium. A basal level of T7 RNA polymerase is expressed in JM 109 (DE3) even in the absence of IPTG, leading to some constitutive expression of the gene of interest. Certain proteins, even when expressed at low levels,
116
VECTORS FOR EXPRESSING CLONED GENES
[8]
may be toxic to the host. To overcome this problem, BL 21 (DE3) pLysE can be used as host strain. Due to the additional pLysE vector this strain produces large amounts of T7 lysozyme, which will bind and inhibit all T7 polymerase produced at a low constitutive level. On IPTG induction, T7 RNA polymerase expression exceeds by far the amounts that are inhibited by the lysozyme present. This leads to transcription of the cloned DNA, which is under the control of the T7 gene 10 promoter in pGEMEX. The presence of pLysE has the further advantage of facilitating the preparation of cell extracts. T7 lysozyme expressed in BL 21 (DE3) pLysE is unable to pass through the inner membrane and therefore cannot disrupt the E. coli cell wall. Treatments that damage the inner membrane (freezethawing or treatment with detergents, e.g., Triton X-100) allow T7 lysozyme to penetrate, causing rapid lysis of cells that contain even small amounts of T7 lysozyme, pLysE confers resistance to chloramphenicol.
Materials and Reagents For Expression of Cloned Polymerase Chain Reaction Products Glycerin stocks for host strains: For long-term storage, 700 #1 of growing or saturated culture is mixed with 100 ~1 of 85% (v/v) glycerol, giving a final concentration of about 10%. Tubes are stored in a - 80° freezer. Bacteria survive for many months in a - 2 0 ° freezer. However, to maintain bacteria for longer periods of time, they should be stored at 80 ° Growth media: LB medium and LB plates are produced according to Sambrook et al. 19Several alternative media are described by Studier ~7 Antibiotics: Antibiotics are added to liquid or solid-growth media with typical concentrations of 20 ~g/ml for ampicillin and 25/zg/ml for chloramphenicol. If agar plates do not contain chloramphenicol, 30 ~1 of a chloramphenicol stock solution (17 mg/ml in ethanol) is plated onto the agar IPTG stock solution: 500 mM in water, sterile filtered Lysis buffer: 50 mM Tris-HCl, pH 8.0, 2 mM NazEDTA, 0.1% (w/v) Triton X-100 Sample buffer, 1 x : 50 mM Tris-HC1, pH 6.8, 2 mM Na2EDTA, 1% (w/v) sodium dodecyl sulfate, 1% (w/v) mercaptoethanol, 8% (v/v) glycerol, 0.025% (w/v) bromphenol blue _
For Quantitation of Polymerase Chain Reaction Products Materials for SDS-polyacrylamide gel electrophoresis (PAGE): See Ref. 19
[8]
cDNA CLONINGBY PCR
117
Coomassie Brilliant blue G-250 (Bio-Rad) Scanning equipment: Laserscan 220 (LKB-Pharmacia, Piscataway, N J) For Recombinant Protein Isolation Reagents for bacterial media, antibiotics, and IPTG: See Materials and Reagents for expression of cloned PCR products Medium for suspension of bacterial culture: 2.4 M sucrose, 40 mM TrisHCI, pH 8.0, 10 mM EDTA Solution for lysis of bacteria: 50 mM N-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid (HEPES)-HCI, pH 8.0, I00 mM KCI, 1 mM EDTA, 1 mM dithiothreitol, lysozyme (100 t~g/ml) (Sigma); dithiothreitol and lysozyme are added just before use Ammonium sulfate: 100% Saturated solution, pH 7.0, adjusted with 5 N NaOH Trichloroacetic acid: 100% (w/v) in H20 Imidazole: 200 mM stock solution, pH 7.0 Dialysis membrane: SpectraPor, No. 3, molecular cutoff Mr 3500 (Spectrum Medical Industries, Los Angeles, CA) DEAE-Sephacel (Pharmacia) Experimental Procedure Protein Expression. Frozen glycerol cultures of JM 109 (DE3) or BL 21 (DE3) pLysE, both containing pGEMEX vector with the insert of interest, are plated on LB agar containing ampicillin or ampicillin and chloramphenicol, respectively. Plates are incubated at 37° to produce single colonies. These colonies are used to inoculate liquid media. Colonies should be stored for no longer than a few days, because of potential damage or loss of pGEMEX during longer periods at 4 °. Single colonies are used to inoculate LB medium containing 20/~g/ml ampicillin as the selective antibiotic [and also 25 pg/ml chloramphenicol if BL 21 (DE3) pLysE is used]. Bacteria are grown at 37° on a shaker. When the culture reaches an OD600 of 0.6-1.5, IPTG is added to a final concentration of 0.4 mM and shaking is continued at 37°. Cells are usually harvested 2-3 hr after induction. This is sufficient for substantial accumulation of recombinant protein and prevents overgrowth of cells that have lost the plasmid or are otherwise unproductive. To stop bacterial growth, the culture is placed on ice followed by purification of the protein of interest. For analysis of the recombinant protein, SDS-PAGE is performed according to Sambrook et al. ~9 In a typical analytical expression experiment bacteria are collected from 1 ml of culture by centrifugation at 5000 g
118
VECTORS FOR EXPRESSING CLONED GENES
18]
for 5 min. The bacterial pellet is resuspended in 100/zl of lysis buffer. After subsequent freezing at - 20° for 30 min and thawing at room temperature for 10 min, 300/xl of 1 × sample buffer is added. Samples are heated for 3-5 min in boiling water and 20 /zl is directly applied to the gel. Generally, this is a sufficient amount for visualizing proteins by staining with Coomassie Brilliant blue (Fig. 3). Notes: The use of freshly grown single colonies from LB plates for expression experiments cannot be overemphasized. Owing to possible instability of the recombinant plasmid it may become lost. Cells that lack the plasmid can rapidly overgrow the culture due to degradation of ampicillin by/3-1actamase that is released into the medium by bacteria containing pGEMEX. If the plasmid is to be maintained in a significant fraction of bacteria, they must never be allowed to grow in the absence of antibiotics. (For a detailed description of problems concerning the loss of plasmids, see Studier et al.17) When the target protein to be expressed may have toxic effects on the host strain, BL 21 (DE3) pLysE should always be used as a host. To maintain the pLysE plasmid, the culture should always be grown in the presence of chloramphenicol. Recombinant Protein Quantitation. Proteins in 20/zl of concentrated E. coli extracts are separated by SDS-PAGE and stained with Coomassie Brilliant blue. The protein concentration of each lane is analyzed by laserscanning densitometry (using Laserscan 220; LKB-Pharmacia). Figures 3 and 4 represent examples of SDS-PAGE and the corresponding scanning profiles of protein extracts from JM 109 (DE3) and BL 21 (DE3) pLysE producing the rat Ca2÷-binding protein parvalbumin. 18 Recombinant Protein Isolation. The procedure for the isolation of recombinant parvalbumin is outlined in Fig. 5. Single colonies of JM 109 (DE3) and BL 21 (DE3) pLysE, both containing the pGEMEX vector with the insert of interest, are used to inoculate 1 liter of LB medium containing 20/~g/ml ampicillin or 20/zg/ml ampicillin combined with 25 /zl/ml chloramphenicol as the selective antibiotics. Bacteria are grown at 37 ° on a shaker. When the culture reaches an OD600 of 0.6-1.5, IPTG is added to a final concentration of 0.4 mM and shaking is continued. Bacteria are harvested 2-3 hr after IPTG induction by centrifugation using a Sorvall (Norwalk, CT) centrifuge (H 6000A rotor) at 4500 rpm. The pellet (approximately 3-4 g) is resuspended in 10 ml of 2.4 M sucrose, 40 mM Tris-HC1, 10 mM EDTA, pH 8.0. The suspension is incubated on ice for 30 min, followed by addition of 40 ml of 50 mM HEPES-HC1, pH 8.0, 100 mM KCI, 1 mM EDTA, 1 mM DTT, and 100/xg/ml lysozyme. Lysis of bacteria is performed at 4° overnight. The lysed bacteria are centrifuged at 47,000 g for 30 min in a Sorvall centrifuge (SS34 rotor), at 4-7 °.
[8]
cDNA CLONINGBY PCR
119
JM 109 (DE3) BL 21 (DE3) pLysE
Mr
A BC
97,400 66,200 45,000 31,000
D E F
AB
C D E F
CoomMsie Gel
21,000 14,400 97,400 66,200 45,000 31,000 21,000
J
Western Blot Analysis
14,400 Fro. 3. Expression of recombinant rat parvalbumin in pGEMEX-PV. Coomassie bluestained acrylamide gels and Western blot analysis of protein extracts are shown for the two E. coli strains JM 109 (DE3) and BL 21 (DE3) pLysE. Lane A, HPLC-purified rat parvalbumin; lane B, low molecular weight standard; lanes C and D, control extracts of host strain without expression vector pGEMEX-PV in the absence (lane C) and in the presence (lane D) of IPTG; lanes E and F, protein extracts of host strain with expression vector pGEMEX-PV encoding rat parvalbumin in the absence (lane E) and in the presence (lane F) of IPTG. Highlevel expression of parvalbumin was observed for both strains. JM 109 (DE3) showed constitutive parvalbumin expression in the absence of IPTG. On IPTG induction parvalbumin production per liter of bacterial culture was further increased and total bacterial protein production was drastically reduced. For BL 21 (DE3) pLysE no parvalbumin expression was detectable with Coomassie staining in the absence of 1PTG. However, using Western blot analysis a faint but parvalbumin-specific signal was observed, indicating some marginal parvalbumin expression in the absence of IPTG. On IPTG induction parvalbumin expression was highly increased,
The clear supernatant is heated rapidly in a boiling water bath at 65-80 ° for 5 min, and then cooled immediately on ice. The denatured proteins are removed by centrifugation at 47,000 g for 30 min at 4-7 °. The supernatant is cooled on ice and 40 ml of 100% saturated ammonium sulfate (pH 7.0) is slowly added to attain a final saturation of 45%. Slow stirring is continued for 1 hr (or overnight). The precipitated proteins are removed by centrifugation at 47,000 g for 15 min at 4-7 °. The parvalbumin-containing supernatant is placed onto an ice water bath and proteins are precipitated by adding slowly a onesixth volume of 100% trichloroacetic acid under constant stirring. Precipi-
120
VECTORS FOR EXPRESSING CLONED GENES + IPTG
- IPTG
Mr
[8]
14,400
45,000
14,400
45,000
I
I
I
I
ILl ~Z W
(5
OZ.
l
JM 109 (DE3) I
IM
0 Q. 4-
> X
IM II/
O Q. t
BL 21 (DE3) pLYS E
> W
÷ __.I!II]IIILI
FIG. 4. Quantitation of recombinant rat parvalbumin expression. SDS gels from Fig. 3 with total bacterial extracts of both JM 109 (DE3) and BL 21 (DE3) pLysE were analyzed using laser-scanning densitometry (laser scan 220, LKB-Pharmacia). JM 109 showed some moderate parvalbumin production of 6-7% in the absence of IPTG and 25-30% in the presence of IPTG with respect to total bacterial protein. In contrast, BL 21 (DE3) pLysE showed no parvalbumin expression in the absence of IPTG, whereas on IPTG induction PV production was 9-12% of total bacterial protein. Using different concentrations of HPLCpurified rat muscle parvalbumin [M. W. Berchtold, C. W. Heizmann, and K. J. Wilson, Anal. Biochem. 129, 120 (1983)] as a standard, JM 109 (DE3) and BL 21 (DE3) pLysE showed a maximal parvalbumin production of 100-140 and 50-70 mg, respectively, per liter bacterial culture.
[8]
c D N A CLONING BY P C R
1 I bacterial culture
A
B C
121
DE
F G
lysis (lysozyme) (over night) supernatant
heat step o (5min., 65-80 C) supernatant
1 ,1
45% (NH~O4 precipitation
supernatant
DEAE-Sephacel 0-200 mM NaCI
parvalbumin immunoreactive peak at 70-80 mM Yield: 25 mg parvalbumin FI6.5. Purification of recombinant rat parvalbumin. In a four-step procedure parvalbumin was purified, taking advantage of its strong heat stability, which is typical for many high affinity hand Ca2+-binding proteins. Trichloroacetic acid precipitation of the 45% saturated ammonium sulfate supernatant was found to be necessary for the binding of parvalbumin to the following DEAE-Sephacel column. Lane C shows the protein composition of total bacterial extract of JM 109 (DE3) pGEMEX-PV after 3 hr of IPTG induction. Lanes D-G represent the protein compositions during subsequent purification steps. Lane A, rat muscle parvalbumin; lane B, low molecular weight standard; lane C, whole bacteria extract; lane D, lysate supernatant; lane E, heat step supernatant; lane F, 45% saturated ammonium sulfate supernatant: lane G, parvalbumin elution peak from DEAE-Sephacel.
tation is continued for 1 hr. The parvalbumin-containing precipitate is collected by centrifugation at 10,000 g (Sorvall GSA rotor) for 30 min, at 4-7 °. The protein pellet is resuspended slowly with 10 ml of 20 mM imidazole buffer (pH 7.0) on a roller shaker at 4° for several hours. The pH is readjusted to pH 7.0 several times with 5 N NaOH. This solution is dialyzed against I00 vol of 20 mM imidazole, pH 7.0, overnight using a SpectraPor dialysis membrane (No. 3, M r c u t o f f 3500). The dialysis buffer is exchanged four times. The dialyzed proteins are loaded onto a DEAESephacel ion-exchange chromatography column (20-ml total volume), preequilibrated with 20 mM imidazole, pH 7.0. For protein elution a linear
122
VECTORS FOR EXPRESSING CLONED GENES
[8]
salt gradient of 0-200 mM NaC1 is used. Parvalbumin immunoreactive fractions are identified using a dot-blot immunoassay :6 with antibodies directed against rat muscle parvalbumin. Parvalbumin elutes at approximately 70-80 m M NaCI. A final yield of 25 mg parvalbumin per liter of bacterial culture is obtained. Notes: Supernatants of subsequent purification steps can be frozen at - 2 0 ° at any time and purification continued at a later time point. Conclusions This chapter describes the detailed methodology for bacterial expression and isolation of heterologous proteins in E. coli. To carry out the described techniques a tissue or cell culture that expresses a gene of interest is required. In addition, a limited amount of sequence information must be available, the minimal requirement being the sequence of a short segment of the protein that allows one to deduce the nucleotide sequence, which is a mixed population of defined sequences owing to codon degeneracy. Initially, the 3' end of a cDNA is amplified by PCR and the PCR product sequenced. This results in a specific cDNA sequence that can be used to generate new oligonucleotides for 5' cDNA amplification. The coding region of the cDNA can then be PCR amplified directly by using the complex mixture of mRNA as a template. Oligonucleotides for this step are designed for convenient cloning directly into the ATG translation start codon of the highly expressed vector gene, which is inducible by IPTG. In conclusion, we present a fast and efficient laboratory protocol for PCR amplification of cDNA, followed by expression, quantitation, and isolation of heterologous proteins, starting with a tissue or cell culture that produces a protein of interest. Acknowledgments This work was supported by Swiss National Foundation Grant 31-28847.90 and Swiss Cancer League Grant 406.89.1. T. M. J. Leu, (Massachusetts Institute of Technology, Cambridge, MA) is acknowledged for advice on PCR as well as for critical reading of the manuscript. We thank M. Ziak and R. Jaussi for their gift of a modified pGEMEX vector and F. W. Studier for several bacterial strains. M. W. B. is supported by a fellowship for independent researcher from the Cloetta Foundation, Zurich.
26 H. Towbin, T. Staehelin, and J. Gordon, Proc. Natl. Acad. Sci. U.S.A. 76, 4350 (1979).
[9]
PREPARATIVE-SCALE
GENE
EXPRESSION
123
[9] G e n e E x p r e s s i o n in C e l l - F r e e S y s t e m o n Preparative Scale
By
VLAD1MIR I. BARANOV a n d A L E X A N D E R S. SPIRIN
Various cell-free systems of protein synthesis and gene expression have been described repeatedly in this series. 1-8 Virtually all of them are based on a crude cell extract (the so-called $30) or a combination of isolated ribosomes and ribosome-free extract (S100) incubated in a test tube with a fixed volume for a short time (up to 1 hr, rarely a few hours). Cell-free translation systems use isolated mRNA or synthetic polynucleotides as templates for proteins or polypeptides to be synthesized. Cellfree coupled transcription-translation systems exploit DNA fragments, plasmids, or isolated genes, which are transcribed by the endogenous R N A polymerase of the bacterial extract coupled with translation of the transcript by ribosomes. Routinely the protein synthesis is recorded by radioactive amino acid incorporation. Specific proteins synthesized can be assayed by their enzymatic or other biological activity, or by reaction with antibodies. The yield is low, so that the systems cannot be used for preparative purposes. Instead of using the batch process in cell-free systems, the continuousflow principle can also be utilized.9 In this case a feeding solution containing amino acids, ATP, and GTP is continuously passed through the incubation mixture, limited by an ultrafiltration membrane. The product is continuously removed from the incubation mixture because it passes through the membrane, while the components of the protein-synthesizing machinery are retained. It proves that the continuous-flow cell-free (CFCF) sysi M. W. Nirenberg, this series, Vol. 6, p. 17. 2 L. M. Gold and M. Schweiger, this series, Vol. 20, p. 537. 3 S. M. Heywood and A. W. Rourke, this series, Vol. 30, p. 669. 4 A. Marcus, D. Efron, and D. Weeks, this series, Vol. 30, p. 749. s W. C. Merrick, this series, Vol. 101, p. 606. 6 E. C. Henshaw and R. Panniers, this series, Vol. 101, p. 616. 7 C. W. Anderson, J, W. Straus, and B. S. Dudock, this series, Vol. 101, p. 635. s H.-Z. Chen and G. Zubay, this series, Vol. 101, p. 674. 9 A. S. Spirin, V. I. Baranov, L. A. Ryabova, S. Y. Ovodov, and Y. B. Alakhov, Science 242, 1162 (1988).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
124
VECTORS FOR EXPRESSING C L O N E D GENES
[9]
tems can work for many hours with constant activity, giving preparative yields of protein products. Reactors
Amicon 8MC Microultrafiltration System The simplest device to realize the continuous-flow process for cellfree translation or coupled transcription-translation is the commercial instrument, Amicon (Danvers, MA) 8MC. This instrument must be placed in a thermostat-controlled box to maintain the temperature of incubation in the ultrafiltration chamber. The incubation mixture containing ribosomes, template polynucleotide, tRNAs, and all necessary protein factors and enzymes is put into the chamber. The type of ultrafiltration membrane, from PM10 or XM30 to YM100 or XM300 (Amicon), is selected depending on the size of polypeptide synthesized, in order to allow it to pass through the membrane. From the beginning of incubation a feeding solution containing substrates (amino acids and nucleoside triphosphates) in a corresponding buffer is pumped into the chamber. The reservoir with the feeding solution is kept cold (in ice or in a refrigerator box nearby). The filtrate containing the polypeptide product is pumped out at the same rate, into a fraction collector placed in the cold. If the volume of the incubation mixture is 1 ml, the flow rate must be 1 to 3 ml/hr. Gentle stirring proceeds during incubation. The process can be continued for 20 to 100 hr. A diagram of the process is shown in Fig. 1. It is important that the incubation mixture be kept sterile. Sterilized water must be used for preparation of all buffers and mixtures. A sterilization filter is recommended on the line from the feeding solution reservoir to the incubation chamber. The presence of NaN3 is desirable in the feeding solution. To prevent oxidation processes from occurring during incubation, the pressure of an inert gas (e.g., He2) can be established in the incubation chamber and in the feeding solution reservoir. This pressure (about 0.5 to 2 additional atmospheres) can be used instead of the force pump. The most serious technical problem in this system is the maintenance of a constant volume of the incubation mixture throughout the process. For this purpose the rates of pumping in and pumping out must be precisely adjusted. In the course of the process, however, the membrane may be partially clogged and thus the rate of filtration can be changed (slowed down), resulting in the influx exceeding the outflux. Therefore, the pumps must be periodically readjusted.
[9l
PREPARATIVE-SCALE GENE EXPRESSION
t Fee~incj s o l u t i o n : i . . Amino•acids." • ATP ...: .'." "GTP " . • Buffer-.". •
.
.
•
~ • "
•
.
.
,
'"".
.
.
•
.
125
.
ULTRAFILTRATION MEMBRANE
: : .i
~V._2
FORCEPUMP
I POM;©
OUTLET
PRODUCT F1G. 1. Scheme of bioreactor for translation based on Amicon 8MC microultrafiltration system.
For technical reasons it is convenient to modify the cap (upper lid) of the incubation chamber in such a way that the feeding solution enters the chamber from the top, not from the bottom.
Upflow Ultrafiltration Column System Another device for CFCF systems is based on a chromatography microcolumn equipped with an ultrafiltration membrane at the outlet and a standard column adaptor at the inlet (Fig. 2). The column is waterjacketed, thus providing easy temperature control. The volume of the incubation mixture in the column is set by the adaptor, so that no air layer exists and no problems with maintaining the volume constancy arise. The column is positioned "upside down," that is, the membrane is up and the adaptor is down. Thus the lighter feeding solution enters the heavier incubation mixture from the bottom and results in convectional intermixing; no stirring is required. The system needs just one pump.
126
[9]
VECTORS FOR EXPRESSING CLONED GENES
Feeding solution:
. ..'i.' . '"
i.
..
' i
Amino acids ATP
GTP
.' • . • .......,
~
! • Buffer .". ,. . i . ] " . " ' " " ' ". ' ~ .I
Ultrafiltration membrane
I~-"-~Cell-free
system
i
['....i....
". '..'.
? ~y_.J PUMP
PRODUCT
FIG. 2. Scheme of bioreactor for translation based on upflow ultrafiltration column system.
Materials and Reagents Plasmid DNA Isolation Plasmid DNA is prepared by the alkaline extraction method ~° with subsequent gel filtration through Sephacryl S-1000. The method based on plasmid DNA precipitation by polyethylene glycol (PEG)-6000 H is also acceptable. Plasmid DNA is stored frozen at - 2 0 ° in TE buffer [10 mM Tris-HC1, pH 8.0, 0.1 mM ethylenediaminetetraacetic acid (EDTA)]. mRNA Preparation Phage. MS2 RNA, brome mosaic virus (BMV) RNA 4, and globin mRNA are isolated according to published procedures, lz-~4Other mRNAs l0 H. C. Birnboim, this series, Vol, 100, p. 243. II R. Treissman, Cell 43, 889 (1985). 12H. O. Voorma, R. Benne, and T. J. A. den Hertog, Eur. J. Biochem. 18, 451 (1971). 13j. D. Stuhbs and P. Kaesberg, Virology 33, 385 (1967). 14A. W. Nienhucs, A. K. Falvey, and W. E. Anderson, this series, Vol. 30, p. 621.
[9]
PREPARATIVE-SCALE GENE EXPRESSION
127
can be prepared by in vitro transcription as described.~5 mRNAs are stored as ethanol precipitates at - 2 0 °. Cell Extract Preparation There are no special requirements for preparation of cell extracts or purified components of cell-free translation systems. Different versions of cell-free extracts may be used. Commercially available extracts are also acceptable. The extracts with a high synthetic capacity, low nuclease activity, and low background of endogenous template translation must be selected. Procedures described in this series may be exploited for preparation of Escherichia coli $30 extracts,8 bacterial 70S ribosomes and SI00 extracts, 2~6 wheat germ extracts, 7 or rabbit reticulocyte lysates) The extracts are stored at - 7 0 °.
Procedures and Results The yield of CFCF systems depends on various factors; such as the synthetic capacity of the cell-free extract used, the expressivity of the template (plasmid DNA, isolated gene, or mRNA), and the solubility of the polypeptide product. Ionic conditions, especially proportions of divalent ( M g 2+) and monovalent (K + and/or NH4 +) ions, are also essential. It should be emphasized, however, that the ionic concentrations used in the examples described below are optimal for each given template; the synthesis of other specific polypeptides in the cell-free systems may require adjustment for optimization. The procedures and results of gene expression in the CFCF systems are illustrated by the following examples. Continuous-Flow Cell-Free Translation Systems In contrast to standard cell-free translation systems, in which saturating amounts of mRNA are generally used, limiting amounts of mRNA in CFCF translation systems are recommended. The excess of mRNA over ribosomes in the incubation mixture may decrease the activity of the system, probably due to degradation of the excess mRNA by endogenous nucleases and competition among RNA fragments for ribosomes. t5 D. A. Melton, P. A, Krieg, M. R. Rebagliati, T. Maniatis, and K. Zinn, Nucleic Acids Res. 12, 7035 (1984). ~6G. H. Chambliss, T. M. Henkin, and J. M. Leventhal, this series, Vol. 101, p. 598.
128
VECTORS FOR EXPRESSING CLONED GENES
[9]
Experiment 1: Synthesis of MS2 Coat Protein in Escherichia coli System 9 Stock solutions Ribosomal buffer (RB): 20 mM Tris-HCl, pH 7.4, 100 mM NH4C1, 10 mM MgCI 2, I mM dithiothreitol (DTT), 0. I mM EDTA (total volume, 100 ml). Store frozen in 5-ml aliquots at - 2 0 ° Salts buffer, 10 × (10 × SB): 160 mM Tris-HCl, pH 7.4, 800 mM NH4 C1, 80 mM MgC1z, 0.8 mM EDTA, 0.2% (w/v) NaN 3 (total volume, 10 ml). Store frozen in 1-ml aliquots at - 2 0 ° Master mixture, 10 x (10 x MM): 50 mM phosphoenol pyruvate (PEP); 10 mM ATP, 2 mM GTP, 8 mM DTT, 0.4 mM folinic acid (total volume, 10 ml). Store frozen in 1-ml aliquots at - 2 0 ° Protein inhibitors mixture, 20 × (20 × PIM): 0.002 mg/ml aprotinin, 0.002 mg/ml leupeptin, 0.002 mg/ml chymostatin (total volume, I ml). Store frozen in 0.05-ml aliquots at - 2 0 ° Mixture (40 x ) of 19 amino acids without leucine (40 × AA-L~u), 1 mM each. Solution of 40 x AA -Leu in water is stored frozen at - 2 0 ° Low-activity radioactive [3H]leucine (LR[3H]Leu), 1.25 mM (specific activity, 0.52 Ci/mmol). Solution of LR[3H]Leu in water is stored frozen at - 20 ° Pyruvate kinase (PK), 1 mg/ml (specific activity, 500 U/rag). Solution in 50% (v/v) glycerol is stored at - 2 0 ° Human placenta ribonuclease inhibitor (HPRI), 25,000 U/ml: Solution in 50% (v/v) glycerol is stored at - 2 0 ° Total tRNA from E. coli (tRNA), 30 mg/ml: Solution of tRNA in water is stored frozen in 0.02-ml aliquots at - 2 0 ° Phage MS2 RNA (MS2): Store in small aliquots as an ethanol precipitate at - 20 °. Before use the RNA precipitate is centrifuged and the pellet is dissolved in water to a final concentration of about 1 mg/ml 70S ribosomes (70S), 32.2 mg/ml: Solution in RB buffer is stored frozen in 0.05-ml aliquots at - 7 0 ° S100 ribosome-free extract from E. coli (S100), 10 mg/ml: Solution in RB buffer is stored frozen in 0.1-ml aliquots at - 7 0 °
Incubation procedure 1. Assemble the bioreactor using an Amicon PM30 ultrafiltration membrane. Adjust the temperature of the reaction chamber to 37°. Wash the reactor and tubings by passing 20 ml RB buffer containing 0.02% (w/v) NaN3 through the whole assembled system at a flow rate of 20 ml/hr.
[9]
129
PREPARATIVE-SCALE GENE EXPRESSION
2. Thaw stock solutions on ice and prepare the incubation mixture and the feeding solution as follows, in the sequence as tabulated below. Stock solutions
I n c u b a t i o n mixture (ml)
F e e di ng solution (ml)
Doubly distilled H20 40 × A A - L e u 10 x MM 10× SB 20 × PIM PK HPRI tRNA RB 70S SI00 LR[3H]Leu MS2 Total v o l u m e
0.416 0.025 0.100 0.100 0.050 0.007 0.002 O.O20 0.050 0.050 0.100 0.020 0.060 1.000
15.15 0.75 3.00 3.00 1.50 ---6,0 --0.60 -30.00
Thus, 1 ml reaction mixture contains 0.6 nmol (1.6 mg) of 70S ribosomes, 1 mg of S100 protein, 0.6 mg of total tRNA, 0.06 nmol (60/xg) of MS2 phage RNA, 7/xg of pyruvate kinase with a specific activity of 500 U/mg, 50 activity units of human placenta ribonuclease inhibitor, and 0.1/xg each of aprotinin, leupeptin, and chymostatin. Both the incubation mixture and the feeding solution contain 20 mM Tris-HCl, pH 7.4, 100 mM NH4CI, 10 mM MgC12, 1 mM DTT, 40/xM folinic acid, 0.1 mM EDTA, 0.02% (w/v) NaN 3 , 1 mM ATP, 0.2 mM GTP, 5 mM phosphoenol pyruvate, 25 /xM [3H]leucine (specific activity, 0.52 Ci/mmol), and 25/~M each of the other 19 amino acids (Table I, experiment 1). 3. Place the reservoir with the feeding solution in the cold (4°) and connect it with the force pump. Fill the pump and tubings with the feeding solution. 4. Keep the incubation mixture for 5 rain at 37° and then put it into the reaction chamber. Pump the feeding solution through the reaction chamber with a flow rate of I ml/hr. 5. Collect the filtrate with a fraction collector (e.g., fractions of 1 ml each) and analyze the fractions. The results of the experiment are shown in Fig. 3. The incorporation of radioactivity into the product synthesized is analyzed by hot trichloroacetic acid precipitation of 0.5-ml aliquots from each fraction. The product is synthesized linearly over 20 hr at 37°. The yield of MS2 coat protein is about 0.1 mg (6 nmol) from 1 ml of incubation mixture, corresponding to
130
VECTORS F O R E X P R E S S I N G C L O N E D G E N E S
[9]
TABLE I COMPOSITION OF INCUBATION MIXTURES AND FEEDING SOLUTIONS FOR C F C F SYSTEMS Experiment Components Incubation mixture only 7 0 S r i b o s o m e s f r o m E. coli (nmol/ml) SI00 ribosome-free extract f r o m E. coli ( m g / m l ) $ 3 0 e x t r a c t f r o m E. coli
1
2
3
4
0.6
.
.
.
.
1.0
.
.
.
.
5
6
m
--
--
--
50
50
(Az60/ml) $30 extract from wheat
--
17
--
--
--
17
e m b r y o s (A260/ml) Rabbit reticulocyte lysate
--
--
200
--
--
--
(A415/ml) T o t a l t R N A f r o m E. coli
0.6
--
--
0.2
0.5
0.06 ----
. 0.1 ---
(mg/ml) MS2 phage RNA (nmol/ml) BMV RNA 4 (nmol/ml) G l o b i n m R N A (/xg/ml) Plasmid pDF34 DNA with gene encoding DHFR (/zg/ml) Plasmid pSP65 DNA with gene encoding DHFR (/xg/ml) Plasmid pSP65 DNA with gene encoding CAT (/zg/ml) SP6 RNA polymerase [ ( U / m l ) × 10 -3] Creatine kinase (U/ml) Pyruvate kinase (U/ml) HPRI (U/ml) A p r o t i n i n (/xg/ml) C h y m o s t a t i n (/xg/ml) L e u p e p t i n (/~g/ml) P e p s t a t i n (/xg/ml) Feeding mixture HEPES, pH 7.6 (mM) T r i s - a c e t a t e , p H 8.2 ( m M ) Tris-HC1, pH 7.4 (mM) NH4CI (mM) Potassium acetate (raM) CaCI2 ( r a M ) Magnesium acetate (mM) MgC12 ( m M )
.
.
. -6.0 --
.
.
.
.
.
-3.5 50 0.1 0.1 O. 1 --
.
. 22.5 . 50 0.1 . O. 1
.
.
40 -. . 112
--10
--
17.5 . 50 -. --
100
2
2
2
--
--
22.5
17.5
---
25
25 0.1
25 0.1
0.1 0.1
0.1 0.I
. 0.1
. --
O. 1
--
0.1
25 --
-55
50 --
40
25
76
100
112
100
. .
. .
150 --
9.5
1.0
.
50
--
. .
1.9
.
----
.
.
0.1
--20 100 --
--150
100
.
400
.
.
.
7
.
11
-14
1.5
1.5
.
(continued)
[9]
PREPARATIVE-SCALE GENE EXPRESSION
13 l
TABLE I (continued) Experiment Components Feeding mixture (continued) Spermidine ( m M ) DTT EDTA
(raM) (mM)
Folinic acid Glycerol [ % PEG-6000
I
2
3
4
5
6
7
--
0.25
0.60
--
0.05
0.25
0.60
1.0
6.0
1.2
4.0
4.0
4.0
0.1
--
--
--
0.1
--
--
--
--
1.5
(/xM)
40
--
--
10
20
(v/v)]
--
1.5
--
--
--
[% (w/v)]
--
--
--
1.6
--
1.5
--
--
--
Hemin ( / z M ) NaN3 ( % ) ATP ( r a M ) CTP ( m M ) GTP ( r a M )
--
--
20
--
--
--
20
0.02
--
--
0.02
0.02
--
--
1.0
2.0
1.0
0.8
1.0
1.0
1.0
--
--
--
0.8
0.4
0.4
0.4
0.2
0.05
0.2
0.8
0.4
0.4
0.4
UTP
--
--
--
0.8
0.4
0.4
0.4
6.0
10.0
(mM)
Creatine phosphate ( m M ) Phosphoenol pyruvate ( m M ) 1 9 amino acids ( - L e u ) (/zM) Low-labeled [ 3 H ] L e u ( / z M ) 1 9 amino acids ( - M e t ) (/zM) Low-labeled [ 3 5 S ] M e t ( / x M )
--
8.0
10.0
--
--
5
--
--
5
10
--
--
25
25
--
100
--
--
--
25
25
--
--
--
--
--
40
--
40
40
40
--
--
40
--
40
40
40
100
about 100 copies of the polypeptide synthesized per copy of MS2 RNA (Table II, PM30). Polyacrylamide gel electrophoresis analysis 17 of the filtrate reveals that the phage coat protein, stained by Coomassie G250, is a predominant polypeptide band. Experiment 2: Synthesis of Brome Mosaic Virus Coat Protein in Wheat Germ System 9 Similar results are obtained with brome mosaic virus (BMV) RNA 4 translation in the wheat embryo extract (Fig. 4). All procedures are the same as described in experiment 1. Composition of the incubation mixture and feeding solution is summarized in Table I, experiment 2. The reaction chamber is equipped with an Amicon XM50 ultrafiltration membrane. One milliliter of incubation mixture is used. The flow rate of the feeding solution is 1 ml/hr at 27 ° (Table II, XM50). Protein synthesis is monitored over 20 hr. The total protein yield is about 0.2 mg (I0 nmol). The product is visualized by Coomassie G250 staining as a predominant 20-kDa polypeptide band in the gel electrophoresis slab. 17 U. K. Laemmli, Nature (London) 227, 680 (1970).
--
132
VECTORS FOR EXPRESSING CLONED GENES
[9]
7.5 o
E "" ~E
I00
•a ~. 5.o
g8 ._. 09
50
0
I
I
I0
20
Time,
hrs
FIG. 3. Kinetics of MS2 coat protein synthesis in the E. coli CFCF translation system. (Inset) Electrophoretic pattern of translation products.
Experiment 3: Synthesis of Globin in Rabbit Reticulocyte System 18 The application of the same flow principle for the translation of isolated globin mRNA in the commercially available micrococcal nuclease-treated rabbit reticulocyte lysate (Amersham, Buckinghamshire, England) was previously demonstrated. ~8 Composition of the incubation mixture and feeding solution is presented in Table I, experiment 3. The reaction chamber is equipped with an Amicon XM100 ultrafiltration membrane. Onehalf milliliter of incubation mixture is used. In this experiment the flow rate is increased to 3 ml/hr. This results in a protein yield of I00 nmol or 2 mg of the product from a 0.5-ml incubation mixture after I00 hr (Table II, XM100). In this experiment Coomassie staining of the gel could not give quantitative information due to an excess of endogenous globin. Therefore, after electrophoresis the synthesized product is visualized by autoradiography. Again, only one main band corresponding to globin contains radioactive material. r8 L. A. Ryabova, S. A. Ortlepp, and V. I. Baranov, Nucleic Acids Res. 17, 4412 (1989).
[9]
133
PREPARATIVE-SCALE GENE EXPRESSION
"5 E
,oilo -t
;2oo
20.I
4.4
--
"~ ca. o3
I00 o O_
0
I0
0
Time, hrs FIG. 4. Kinetics of BMV coat protein synthesis in the wheat germ CFCF translation system. (Inset) Electrophoretic pattern of translation products.
Continuous-Flow Cell-Free Coupled Transcription-Translation System (with Endogenous RNA Polymerase) The CFCF translation systems described above require isolated individual mRNAs. An alternative approach makes use of the coupled transcription-translation in a cell-free extract. Bacterial extracts are known to contain a DNA-dependent R N A polymerase necessary for transcription. The application of the flow-through principle to the coupled transcription-translation system using the endogenous RNA polymerase and exogenous DNA (gene or plasmid) has the advantages of the CFCF translation systems, resulting in the production of preparative amounts of proteins. A system of this kind, however, is limited by the use of bacterial extracts.
Experiment 4: Synthesis of Dihydrofolate Reductase and fl-Lactamase in Escherichia coli System 19 Figure 5 demonstrates the DNA-directed synthesis of both/3-1actamase (Bla) and dihydrofolate reductase (DHFR) in the CFCF coupled transcription-translation system based on the nucleic acid-free E. coli $30 extract. 19 V. I. Baranov, I. Y. Morozov, S. A. Ortlepp, and A. S. Spirin, Gene 84, 463 (1989).
134
[9l
VECTORS FOR EXPRESSING CLONED GENES T A B L E II MEMBRANES, PARAMETERS, AND YIELDS IN C F C F SYSTEMS Amicon ultrafiltration membrane
Parameter
PM30
XM50
Volume of incubation mixture (ml) Temperature (°C) Time of operation (h) Flow rate (ml/hr) Yield (mg) MS2 phage coat protein BMV R N A 4 coat protein Globin (/3-Lactamase + D H F R ) DHFR CAT Efficiency of system (ttg product) per hour per milliliter of incubation mixture Specific activity (U/~mol)
1.0
1.0
0.5
37 20 1.0
27 20 1.0
30 100 3.0
0. l ---. . 5
. 0.2 ---
.
.
XM100
.
. 10
.
.
YM100
XM300
1.0
0.5
0.5
0.5
37 50 2- 3
37 24 1.5
24 24 1.5
34 35 1.5
. .
. .
. 0.22 .
40
.
YM100
. . 2.0 -.
.
XM100
. 4.4
.
. .
.
. .
-0.011
. -0.09
0.9
7.5
.
14
--0.06 3.5
25
Plasmid pDF34, carrying the two corresponding genes, s° is expressed for 50 hr at 37°. The reaction chamber is equipped with an Amicon XM100 ultrafiltration membrane (Table II, XM100). One milliliter of incubation mixture is used. The feeding solution containing all 4 nucleoside triphosphates and 20 amino acids (Table I, experiment 4) is passed through the incubation chamber with a variable flow rate. The synthesis rate directly depends on the flow rate: the switch from 3 to 2 ml/hr decreases the synthesis rate more than twofold, and the subsequent switch from 2 to 3 ml/hr restores the synthesis rate. The total protein yield is more than 0.2 mg. Only two bands corresponding to fl-lactamase and dihydrofolate reductase are detected in approximately equimolar amounts by polyacrylamide gel electrophoresis and subsequent fluorography.
Continuous-Flow Cell-Free Transcription-Translation Systems with Phage (SP6 or T7) RNA Polymerase There are several limitations in the use of the endogenous RNA polymerase of bacterial extracts. Most of them are caused by problems of transcriptional regulation, proper promoters, termination of transcription, 20 N. V. Murzina and A. T. Gudkov, Protein Eng. 3, 709 (1990).
135
P R E P A R A T I V E - S C AGENE L E EXPRESSION
[9]
300 ¢0 o
E~ 4
200
N O
IO0 13..
"5 i
t3 l/h o
I
I
I
20
40
60
Time, hrs FIG. 5. Kinetics of expression of/3-1actamase and dihydrofolate reductase-encoding genes in the E, coli CFCF system with endogenous RNA polymerase. (Inset) Electrophoretic pattern of translation products.
and so on. As a rule, a complete circular plasmid with the gene of interest, a selectable marker gene, and all the necessary regulatory elements are involved. The selection gene is also transcribed and translated. For example, in experiment 4 the gene of interest (fol), encoding dihydrofolate reductase, was expressed together with the selection gene (bla) coding fl-lactamase. In addition, as mentioned above, such a system is applicable only to bacterial extracts. The next experiments demonstrate that these limitations may be overcome by the use of the CFCF transcription-translation systems with an exogenous bacteriophage RNA polymerase instead of an endogenous cellular RNA polymerase. Phage SP6 or T7 RNA polymerases are the most widely used now for the in vitro synthesis of mRNA by transcription from plasmid DNA or synthetic genes. A set of commercially available vectors containing promoters for these polymerases, such as pSP65, pSP73, or pGEM, have been constructed from E. coli plasmids. It has been found that both T7 and SP6 polymerases can efficiently transcribe mRNA from the DNA plasmid in the CFCF systems, utilizing both pro-
136
VECTORS FOR EXPRESSING CLONED GENES
[9]
karyotic and eukaryotic extracts (V. I. Baranov, L. A. Ryabova, O. B. Yarchuck, and A. S. Spirin, 1990, unpublished observations). Hence, instead of preparing mRNA one can use the DNA plasmid directly for preparative gene expression in the CFCF systems based on cell-free extracts from different types of cells.
Experiment 5: Synthesis of Dihydrofolate Reductase in Escherichia coli Continuous-Flow Cell-Free Transcription-Translation System Plasmid pSP65 containing the bla gene under the control of the E. coli RNA polymerase promoter and DHFR gene under the control of the SP6 promoter, kindly provided by Drs. N. V. Murzina and A. T. Gudkov (Institute of Protein Research, Russian Academy of Sciences). The reaction chamber is equipped with an Amicon YM100 ultrafiltration membrane. A 0.5-ml aliquot of incubation mixture containing the SP6 RNA polymerase is used. The feeding solution contains all 4 nucleoside triphosphates, 20 amino acids, and rifampicin, the inhibitor of bacterial RNA polymerases (Table I, experiment 5). The flow rate of the feeding solution is 1.5 ml/hr at 37 °. The protein synthesis is monitored over 24 hr (Fig. 6). The total protein yield is about 0.01 mg (0.6 nmol). One main band corresponding to DHFR may be detected by electrophoretic analysis with subsequent autoradiography. In parallel, the enzyme activity is measured 2a as well. The specific activity of DHFR is about 14 U//xmol (Table II, YM100).
Experiment 6: Synthesis of Dihydrofolate Reductase in Wheat Germ Continuous-Flow Cell-Free Transcription-Translation System Figure 7 illustrates the expression of the plasmid pSP65 containing the gene encoding DHFR under the control of the SP6 promoter in the wheat embryo extract supplemented with SP6 RNA polymerase (Table I, experiment 6). The reaction chamber is equipped with an Amicon YM100 ultrafiltration membrane. A 0.5-ml aliquot of incubation mixture is used. The flow rate of the feeding solution is 1.5 ml/hr at 24° (Table II, YM100). Protein synthesis is monitored over 24 hr. The total protein yield is about 0.09 mg (5 nmol). One main band corresponding to DHFR is detected by electrophoretic analysis with subsequent autoradiography. In parallel, the enzyme activity is measured 21 as well. The specific activity of DHFR is about 25 U/tzmol. 2J D. P. Baccanari and S. S. Joyner,
Biochemistry 20, 1710 (1981).
[9]
137
PREPARATIVE-SCALE GENE EXPRESSION
kOa 97.468,0
-
0.6
10 DHFR
o - -
184 -
~
t~ .-a
0.4
EL
0.2
I
I
I
5
I0
15
,
I
20
Time, hrs FIG. 6. Kinetics of expression of dihydrofolate reductase-encoding gene in the E. coli CFCF system with bacteriophage SP6 RNA polymerase. (Inset) Electrophoretic pattern of translation products.
Experiment 7: Synthesis of Chloramphenicol Acetyltransferase in Rabbit Reticulocyte Continuous-Flow Cell-Free Transcription-Translation System In this experiment chloramphenicol acetyltransferase (CAT) is synthesized in a rabbit reticulocyte lysate supplemented with SP6 RNA polymerase (Fig. 8). Composition of the incubation mixture and feeding solution is listed in Table I (experiment 7). The reaction chamber is equipped with an Amicon XM300 ultrafiltration membrane. A 0.5-ml aliquot of incubation mixture is used. The flow rate of the feeding solution is 1.5 ml/hr at 34° (Table II, XM300). The protein synthesis is monitored over 35 hr. Total protein yield is about 0.06 mg (2.5 nmol). The enzyme activity is measured 22 in parallel.
22 W. M. Hodges and D. E. Hruby, Anal. Biochem. 1611, 65 (1987).
138
VECTORS FOR EXPRESSING
CLONED
[9]
GENES i
'100
4
4°= ~'C3
50
03 03
o~ -',1
No z
._= -~
03 >..
eg
D..
l
r
5
I0
I
15
I
20
Time, hrs
FIG. 7. Kinetics of expression of dihydrofolatereductase-encodinggene in the wheat germ CFCFsystemwithbacteriophageSP6RNApolymerase.(Inset)Electrophoreticpattern of translation products.
Discussion and Remarks
Advantages It is clear that the first and main advantage of the CFCF systems is their long lifetime, as compared with standard cell-free systems. The CFCF systems display linear kinetics of polypeptide synthesis over tens of hours. This provides preparative yields of corresponding polypeptides and proteins. According to our experience with the CFCF systems, the yield can vary from 50/zg to 4 mg of polypeptide or protein from 1 ml of incubation mixture depending on the molecular mass of the product, its solubility, expressibility of template, type and quality of the cell extract, and so on. Another important advantage of CFCF systems of gene expression is the relative purity of the polypeptide or protein product. Indeed, in this case the product is in the filtrate but not in a complex mixture of the cell extract; the only contaminants are cell extract proteins leaking out from the incubation mixture. With highly expressible mRNA the purity of the
[9]
PREPARATIVE-SCALE GENE EXPRESSION
12
Q) ~0
kcChl - -
Chl t-
139
I B
/
~50 t~
I
--
0
~
¢/) t--
oE
f 0
I
1
IO
O
25
I
t
20
L
I
50
Time, hrs F i t . 8. Kinetics of expression of chloramphenicol acetyltransferase-encoding gene in the rabbit reticulocyte CFCF system with bacteriophage SP6 RNA polymerase. (Inset) Thinlayer chromatography of chloramphenicol acetyltransferase assay products. Functional activity of protein synthesized in the CFCF system after (1) 0 hr of incubation; (2) 12 hr of incubation.
product in the filtrate can reach 80%, especially if the fractions collected during the first few hours are discarded. For instance, 20 hr of expression of interleukin 4 mRNA in an experimental large-scale reactor with the use of wheat germ extract gave 50 mg of the protein with 80% purity, in 1 liter of the filtrate (Y. B. Alakhov, S. Y. Ovodov, and L. M. Vinokurov, Institute of Protein Research, Russian Academy of Sciences unpublished observations, 1990).
Unexplained Features of Continuous-Flow Cell-Free Systems The flow rate of the feeding solution through the incubation mixture is found to be the most critical factor in maintaining the system activity for a long period of time. It is not clear yet whether the continuous removal of inhibitory low molecular mass products or their derivatives (e.g., nucleotides and phosphates), or the elimination of some regulatory elements of
140
VECTORS FOR EXPRESSING CLONED GENES
[9]
the cell extract, or the prevention of a feedback inhibition of ribosomes by newly synthesized polypeptides is responsible for this effect. Another interesting feature of CFCF systems is the relative stability of RNA templates during many hours of incubation at physiological temperatures. At least in the cases of CFCF translation systems (Figs. 3 and 4), mRNA was present in a limiting amount, so that any degradation of mRNA could be seen from the change of the kinetic curve slope. Thus, m R N A seemed to be resistant against phosphodiesterase activities present (although at a low level) in all the types of extracts used. Our explanation of this phenomenon is that mRNA in an actively working system is protected by translating ribosomes. The most striking observation with CFCF systems is the absence of transmembrane leakage, or very slow leakage, of tRNAs and low molecular mass proteins involved in translation. The systems work linearly during tens of hours even when the Amicon ultrafiltration membrane of the YM100 type is used. This membrane is permeable to tRNAs, prokaryotic IF-1 and IF-2, some eukaryotic IFs, as well as, to a less extent, prokaryotic EF-T,, EF-Ts, RF, and eukaryotic eEF-1L and eRF. EF-G, eEF-2 and several aminoacyl-tRNA synthetases should be added to this list in the case of the XM300 membrane. We believe that tRNAs and all other proteins listed above are retained in the translation mixture due to their participation in dynamic complexes with other components of the proteinsynthesizing machinery, so that they virtually do not exist as free individual molecules. Problems and Limitations
With weakly expressed messages, however, contaminating proteins are still a problem in CFCF systems. Leakage of many proteins from the incubation mixture continues for a long time, so that the product polypeptide cannot be seen as a predominant band in electrophoresis during the first 6 to 20 hr of incubation. In this case radioactive amino acid incorporation is required to identify the protein synthesized among other bands in an electrophoretic pattern. This problem can be partially solved by the better choice of an ultrafiltration membrane and, more radically, by the use of a better (more expressible) construct of the corresponding gene or mRNA. Unfortunately, construction of highly translatable messages is not a simple and predictable matter. Many viral RNAs are known to be strong messages. These RNAs are also effectively expressed in CFCF translation systems (see, e.g., Figs. 3 and 4). Viral RNA leaders can be used for
[9]
PREPARATIVE-SCALE GENE EXPRESSION
141
construction of chimeric messenger RNAs containing a required sense sequence to be translated in vitro. 23-25 The CFCF systems are seriously limited by the use of cell extracts with low ribonuclease activities. That is why only the $30 extracts of wheat germs and rabbit reticulocytes and the SI00 extract ofE. coli MRE600 have been used up to now. High quality of the extracts is a strict prerequisite for successful experiments with CFCF systems. The $30 extract of E. coli MRE-600 gives nonreproducible results in CFCF translation systems, presumably because of ribonuclease damage to mRNA. Exonuclease-proof constructs of mRNAs seem to be possible and demanded. A search for new universal inhibitors of ribonucleases is an important task. The specific problem of CFCF systems is an inevitably high dilution of the product in the filtrate. Concentration by subsequent ultrafiltration is recommended; at the same time the use of the concentrating ultrafiltration with a properly chosen membrane provides the removal of smaller contaminating proteins. The other ways to concentrate and purify the product are salting out with ammonium sulfate, or specific adsorption and subsequent elution. Immunoadsorbents or affinity columns can be used in line with the reactor. Some newly synthesized proteins have a tendency to aggregate, either directly in the incubation mixture or on the surface of the ultrafiltration membrane. In both cases it results in clogging of the membrane pores and slowing down of the flow, up to a full stop. The kinetic curve deviates from linearity and goes down. Protein aggregation and pore clogging were the main reason why the synthesis of viral coat proteins was not continued after 20 hr (Figs. 3 and 4). Sometimes the substitution of a new filter for the clogged one in the course of incubation helps and does not change the linearity of the kinetic curve. On the whole, the aggregation may be a significant problem in the cell-free synthesis of subunit, large, and hydrophobic proteins. Perspectives
First of all, the CFCF systems have undoubtedly good prospects for use in the preparative syntheses of polypeptides and proteins that are poorly expressed in vivo. This involves primarily cytotoxic products. 23 S. A. Jobling and L. Gehrke, Nature (London) 325, 622 (1987). 24 D. R. Gallie, D. E. Sleat, J. W. Watts, P. C. Turner, and T. M. A. Wilson, Nucleic Acids Res. 15~ 3257 (1987). z5 S. K. Jang, M. D. Davies, R. J. Kaufman, and E. Wimmer, J. Virol. 63, 1651 (1989).
142
VECTORS FOR EXPRESSING CLONED GENES
[9]
Furthermore, polypeptides and proteins unstable in the cell can be produced better under cell-free conditions. One of the scientific applications is the synthesis of protein-folding intermediates. 26 Due to the fact that the product is filtered out of the incubation mixture and can be visualized as the only newly synthesized peptide, identification and isolation of expression products of open reading frames is possible, even without knowing their properties and functions. For the same reasons, CFCF systems seem to be ideal for the synthesis of theoretically designed proteins. 27 The long-term process offered in CFCF systems and the flux principle used open new possibilities for studies of transcriptional and translational regulation. In particular, various effectors can be added to the incubation mixture for a short time (e.g., as a pulse in the flowing feeding solution) to follow the prolonged aftereffects. One of the most promising applications of CFCF systems will be in vitro protein engineering. Synthetic or isolated site-specifically altered genes can be directly expressed, and all products, independent of their functional and physical properties, can be visualized and investigated. With the use of artificial (anticodon changed) and misacylated tRNAs, protein variants may be obtained without site-specific mutagenesis at the level of genes, but through in vitro site-specific incorporation of different amino acids. Using the same approach, site-specific incorporations of unnatural and labeled amino acids into proteins is possible. 28 Realization of co- and posttranslational modifications of newly synthesized proteins in cell-free conditions is one of the most important and worthy tasks. In any case, involvement of intracellular membrane fractions in the cell-free translation process must be considered for the future.
26 A. N. Fedorov, B. Friguet, L. Djavadi-Ohaniance, Y. B. Alakhov, and M. E. Goldberg, J. Mol. Biol. in press. (1992). ~7 A. N. Fedorov, D. A. Dolgikh, V. V. Chemeris, B. K. Chernov, A. V. Finkelstein, A. A. Schulga, Y. B. Alakhov, M. P. Kirpichnikov, and O. B. Ptitsyn, J. Mol. Biol. 225, 927 (1992). 28 C. J. Noren, S. J. Anthony-Cahill, M. C. Griffith, and P. G. Schultz, Science 244, 182 (1989).
[10]
INSECT VIRUS TRANSLATIONAL INITIATION SIGNAL
143
[10] C e l l - F r e e E x p r e s s i o n V e c t o r : U s e o f I n s e c t V i r u s T r a n s l a t i o n a l I n i t i a t i o n S i g n a l for in Vitro G e n e E x p r e s s i o n
By
B I M A L E N D U DASMAHAPATRA
Highly efficient in vitro transcription of DNA, L2 coupled with cellfree translation of transcribed RNAs into protein, 3'4 greatly enhances the ability to express and study genes outside living cells. 5'6 Gene expression in a cell-free system avoids various cellular regulatory mechanisms that often pose serious problems for in vivo gene expression. Some of these problems are (1) rapid degradation of the expressed gene products by host enzymes, 7 (2) the adverse effect of the foreign protein on the normal host metabolism, and (3) sensitivity of the heterologous RNA transcripts to various cellular regulatory controls, such as splicing and transportation into the cytoplasm. In addition, cell-free gene expression makes it possible to study the expressed gene products without their prior purification from host cell proteins of similar structure and function. 6 Cell-free expression is limited, however, to the use of genes that carry their own signals to initiate translation efficiently. Translational initiation signals are contained within sequences in the 5' untranslated region (UTR) as well as in the vicinity of the initiator ATG ofa eukaryotic mRNA, where ribosomes bind and function to initiate protein synthesis. 8 Eukaryotic mRNAs, unlike prokaryotic mRNAs, do not have a consensus ribosomebinding sequence and show enormous variation in their 5' UTR in both sequence and in length. 9 This variability in translational initiation signals may result in differences in the quantity of protein molecules produced in a given cell-free expression system. To overcome these problems we have cloned a short duplex oligonuclet D. A. Melton, P. A. Krieg, M. R. Rebagliati, T. Maniatis, K. Zin, and M. R. Green, Nucleic Acids Res. 12, 7035 (1984). 2 E. T. Butler and M. J. Chamberlin, J. Biol. Chem. 257, 5772 (1982). 3 p. A. Krieg and D. A. Melton, Nucleic Acids Res. 12, 7057 (1984). 4 M. Spiess and H. Lodish, Cell 44, 177 (1986). s D. S. Gill, D. Chattopadhyay, and A. K. Banerjee, Proc. Natl. Acad. Sci. U.S.A. 83, 8873 (1986). 6 G. D. Parks, G. M. Duke, and A. C. Palmenberg, J. Virol. 60, 376 (1986). 7 K. Talmadge and W. Gilbert, Proe. Natl. Acad. Sci. U.S.A. 79, 1830 (1982). 8 M. Kozak, Cell 22, 7 (1980). 9 M. Kozak, Microbiol. Rev. 47, 1 (1983).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
144
VECTORS FOR EXPRESSING CLONED GENES
[10]
otide, which contains the 5' UTR and the initiator ATG of black beetle virus (BBV) RNA 1,1°into the plasmid pGEM2 (Promega Biotec, Madison, WI) to construct the cell-free expression plasmid pBD7 (Fig. I). 11 The unique restriction sites immediately after the BBV sequence provide multiple cloning sites for inserting genes or gene segments to be expressed. The chimeric RNAs produced in cell-free transcription will, therefore, have a built-in BBV translational initiation signal. As an example of the general utility of the vector, we describe here the use of the BBV signal to direct the cell-free expression of a biologically active coxsackievirus (CV) 3C protease.
Materials and Methods Bacterial Strains Escherichia coli strain JM 101 is used in all cloning and plasmid DNA amplification experiments. Enzymes, Plasmids, and Radiochemicals
All restriction endonucleases as well as DNA-modifying enzymes are obtained from New England BioLabs (Beverly, MA). The plasmid, pGEM2, the rabbit reticulocyte cell-free translation kit, T7 RNA polymerase, and RNasin may be purchased from Promega Biotech. The cap analog, GpppG, and the ribonucleotides are obtained from Pharmacia (Piscataway, NJ). L-[35S]Methionine (I 100 Ci/mmol) is supplied by Amersham (Arlington Heights, IL). The recombinant DNA methods, which include (1) digestion of DNA with various restriction endonucleases, (2) modification of the ends of DNA by Klenow fragment of E. coli DNA polymerase, and (3) joining of DNA molecules by T4 DNA ligase, are performed according to the recommendation of the suppliers. Agarose gel electrophoresis to separate DNA molecules of different sizes and electroelution of DNA molecules from agarose gel slices are carried out according to the procedures described by Maniatis et al. 12
10 B. Dasmahapatra, R. Dasgupta, A. Ghosh, and P. Kaesberg, J. Mol. Biol. 182, 183 (1985). ii B. Dasmahapatra, E. J. Rozhon, and J. Schwartz, Nucleic Acids Res. 15, 3933 (1987). ~2T. Maniatis, E. F. Fritsch, and J. Sambrook, "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982.
[10]
INSECT
VIRUS
TRANSLATIONAL
INITIATION
145
SIGNAL
! !
;> o L )
u~ e
g2~ ~
m
0
--
~ ~.~ [.. ~
.g<
me~
o~
]~
u~8 I
N
. "~.
-
~X"
•
~
~
I ~Sd ._~ <
146
VECTORS FOR EXPRESSING CLONED GENES
[10]
Construction of Plasmid DNAs Expressing Native and Mutant 3C Protease o f Coxsackievirus The plasmid pCB 111 35 (Fig. 2) encoding the 3C protease and flanking regions of the coxsackievirus B3 polyprotein ~3is digested with XbaI and AvaI and the 1826-bp DNA fragment containing the viral sequence from nucleotide 4947 to nucleotide 6773 TMis gel purified. To construct a CV 3C protease expression plasmid including the BBV translational initiation signal, the gel-purified viral DNA fragment is cloned into XbaI/AvaIcut pBD7 (Fig. 2). Thus the resulting plasmid, pC1B1, contains the CV sequence encoding the carboxy-terminal 26 amino acids of 2C, all of 3A, 3B, and 3C, and the amino-terminal 288 amino acids of 3D, fused in-frame with the BBV initiator ATG. The plasmid pC11B9, which expresses a mutant 3C protease, is constructed as follows: the plasmid pC 1B 1 is linearized with Bali restriction endonuclease and ligated to a 12-base pair (bp) oligonucleotide linker (GGATCCGGATCC; New England BioLabs). Instead ofphosphorylating the linker prior to its ligation to the linearized pC1B 1 DNA, nonphosphorylated oligonucleotide linker is ligated, in a 100 M excess with the gelpurified Bali-cut pC 1B 1 DNA, using T4 DNA ligase. Because the linkers are not phosphorylated, they cannot self-ligate, which ensures that a single copy of the linker molecule is inserted. Cell-Free Transcription of Plasmid DNAs Plasmids pC1BI and pCl IB9 are linearized with AvaI and purified by extraction with phenol/chloroform and precipitation with ethanol prior to being used in transcription reactions.~5 Transcription Protocol l. The reaction mixture contains the following: 10/~1 5 × transcription buffer (200 mM Tris-HC1, pH 7.5, 30 mM MgCI2, 10 mM spermidine, 50 mM NaC1) 5/xl 4 mM ATP, CTP, UTP (prepared by mixing equal amounts of rNTP stocks) 5/xl 2 mM GTP 5/xl 10 mM GpppG 1/~1 500 mM dithiothreitol (DTT) I3 S. Tracy, N. M. Chapman, and H. L. Liu, Arch. Virol. 85, 157 (1985). I4 A. M. Lindberg, P. K. Stalhandske, and U. Pettersson, Virology 156, 50 (1987). ~5 M. M. Konarska, R. A. Padgett, and P. A. Sharp, Cell 38, 331 (1984).
[10]
INSECT
VIRUS
TRANSLATIONAL
INITIATION
147
SIGNAL
e~ ~D
I~^V
-V-] J I
I
i
i
I
r..)
C O ,°,
u~ I~qX e-, x
"4
-0) e-
IlSd
e~
'
~
03
" ro ¢q
I~S-
~
o
7~ ,-
o
I
IOO S
>P
l~^V/l~m S -iHtUg 8 I ~qx
;>
1 lSd
t~ r~ c~
148
VECTORS FOR EXPRESSING CLONED GENES
[10]
2/.d linearized plasmid DNA (1 /zg//zl) in water or in Tris-ethylenediaminetetraacetic acid (EDTA) buffer RNasin [final concentration = 1 unit (U)/tzl] T7 RNA polymerase (10 U) Bring the volume up to 50/xl with diethyl pyrocarbonate (DEPC)-treated water. Each of the components should be added in the order shown and the mixture should be kept at room temperature during the addition of each successive component, because DNA can precipitate in the presence of spermidine if kept at 4 °. Nucleotide stock solutions should be neutralized to pH 7. 2. Incubate at 37 ° for 60 min. 3. Add RNase-free DNase to a concentration of ! U//zg DNA. 4. Incubate for 15 min at 37°. 5. Extract with an equal volume of phenol-chloroform. 6. Add 0.5 vol 7.5 M ammonium acetate and 2.5 vol ethanol to precipitate the RNA. 7. Spin, wash the pellet with 70% (v/v) ethanol, and dry the pellet. 8. Resuspend the RNA in 20/.d DEPC-treated water. Transcripts may be analyzed by electrophoresis on an RNase-free agarose gel) 2 Cell-Free Translation of in Vitro-Synthesized RNA Cell-free translations of in vitro synthesized RNA using rabbit reticulocyte lysates are carried out in the presence of [35S]methionine according to the protocol described by Promega Biotec. Translation Protocol 1. Combine the following: 35/xl nuclease-treated reticulocyte lysate (slowly thawed in ice-water) 6 ~l DEPC-treated water 1/A RNasin (40 U//zl) 1/xl 1 mM amino acid mixture (minus methionine) 2/xl RNA transcript (I-2/xg) 15a 5/xl [35S]methionine (1150 Ci/mmol) at 9.2 mCi/ml 2. Incubate at 30° for 60 min. 15a Optimal concentrations for RNA transcripts are determined by translating different amounts of the RNA transcript with the same amount of assay mixture. Also, denaturing RNA transcripts by keeping them at 65° for 5 min prior to translation removes secondary structure and may improve translational efficiency.
[10]
INSECT VIRUS TRANSLATIONAL INITIATION SIGNAL
C
1
149
2
P1 3CD 3D
VP1
3C i
i i
F1G. 3. SDS-PAGE analysis of cell-free protein synthesis from plasmids pC1BI and pC 11B9. Plasmid DNAs were transcribed in vitroand the products were translated in reticulocyte lysates as described in Materials and Methods. Lanes 1 and 2 represent translation mixtures programmed with RNA molecules derived from plasmids pCI1B9 and pC1BI, respectively. Lane C contains 35S-labeled proteins produced in the CVB3-infected HeLa cells.
Analysis o f Coxsackievirus 3C Protease Expression Sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( S D S P A G E ) 16 is used to analyze the CV3C protease expression in reticulocyte lysates. A 2-/A aliquot of [35S]methionine-labeled cell-free translation products is diluted with sample buffer [0.0625 M Tris-HC1, p H 6.8, 2% (w/v) SDS, 5% (v/v) 2-mercaptoethanol, 10% (v/v) glycerol, and 0.001% (v/v) b r o m p h e n o l blue] and heated at 90 ° for 5 rain before loading onto an S D S - p o l y a c r y l a m i d e gel as described in Fig. 3. After electrophoresis the 16U. K. Laemmli, Nature (London) 227, 680 (1970).
150
VECTORS FOR EXPRESSING CLONED GENES
[10]
gel is soaked in 40% (v/v) methanol containing 10% (v/v) acetic acid, dried under vacuum, and exposed to Kodak (Rochester, NY) XAR-5 film. Results The cell-free expression plasmid, pBD7, contains the 5' UTR and initiator ATG of BBV RNA 1 downstream of the promoter for T7 RNA polymerase (see Fig. 2). Therefore, RNA transcripts made from pBD7 derivatives by the bacteriophage RNA polymerase have the BBV translational initiation signal at their 5' termini. The expression plasmid pC1B I contains the coxsackievirus coding sequence fused in frame to the BBV initiator ATG. The chimeric RNA transcribed from the AvaI-linearized pC 1B 1 encodes a protein with a molecular weight of 72,000 that derives the initiator methionine and the second amino acid, threonine, from the vector sequence and the rest from CV polyprotein (Fig. 2). In coxsackievirus-infected cells a single large viral polyprotein (250K) is produced that, in its nascent form, is proteolytically processed by two viral proteases, 2A and 3C, into several mature viral proteins. 17.18The 72K CV protein encoded in the pC 1B 1 transcript contains 26 amino acids from the carboxyl end of 2C, all of 3A, 3B, the protease 3C, and the aminoterminal 288 amino acids of the polymerase 3D. 14The 3C protease autocatalytically cleaves itself out of the viral precursor polyprotcin between glutamine and glycine pairs present 183 amino acids apart, thus creating its own amino and carboxy termini. 17The RNA transcript produced from the plasmid pC 11B9 contains an in-frame insertion of four amino acids (P, D, P, D) within the 3C protease sequence of the same 72K protein encoded by the pC1B1 transcript. The mutant 3C protease encoded by the pC1 IB9 transcript is defective in its autocatalytic activity. 18aChimeric RNAs were translated in rabbit reticulocyte lysates, and synthesis of CV-specific proteins with the BBV translational initiation signal was analyzed by SDS-polyacrylamide gel electrophoresis. The results (Fig. 3) showed that a major protein, with an approximate molecular weight of 72K, was synthesized from pC 11B9 transcripts encoding the mutant 3C protease (Fig. 3, lane 1). As expected, the 72K protein was absent in the translation ofpC1Bl-transcripts (Fig. 3, lane 2). Instead, several other proteins ranging from 20K to 56K, presumably produced by the autocatalytic activity of the 3C protease, were seen (Fig. 3, lane 2). 17 H. G. Krausslich and E. Wimmer, A n n u . Rev. Biochem. 57, 701 (1988). 18 A. C. Palmenberg, in "Positive Strand RNA Viruses" (M. A. Brinton and R. R. Rueckert, eds.), p. 25. New York, 1987. taa B. Dasmahapatra, E. J. Rozhon, A. M. Hart, S. Cox, S. Tracy, and J. Schwartz, Virus Res. 20, 237 (1991).
[10]
INSECT VIRUS TRANSLATIONAL INITIATION SIGNAL
151
The 20K protein comigrates with the native 3C protease produced in CVB3-infected HeLa cells. This indicates that the 3C protease sequence contained in the 72K protein is autocatalytically active. The other prominent protein bands represent the products of the autocatalytic activity in the 72K protein. These results suggest that the BBV translational initiation signal is able to direct synthesis of CVB3-specific protein in vitro. Comments The BBV sequence contained in the plasmid pBD7 is sufficient to direct translational expression of the coxsackievirus coding sequence, which lacks its own signal for the initiation of translation. The expressed gene products are biologically active. The plasmid pBD7 has also been used for the cell-free expression of c-fos and Epstein-Barr virus Zta genes, in which the BBV sequence has been reported to yield at least 20-fold more Zta protein than the translational initiation signal provided by the Zta leader sequence. ~9 Untranslated leader sequences of eukaryotic mRNAs contain information that not only guides ribosomes to initiate protein synthesis correctly, but also regulates the efficiency of translation. Some of these parameters influencing the efficiency of translation of an mRNA are the cap structure at the 5' end of the mRNA, the sequences flanking the initiator ATG codon, the presence or absence of upstream ATG codons, and the secondary structure in the 5' U T R . 9'2°'21 The BBV sequence contained in the plasmid pBD7 is not predicted to form stable secondary structure. Moreover, it contains an adenosine residue at the - 3 position relative to the initiation codon. The chimeric RNAs produced from pBD7 derivatives are capped at their 5' termini. Native viral RNAs are translated efficiently.22'23 These may explain the relative efficiency of the BBV sequence in pBD7 in directing translation of heterologous coding sequences. Acknowledgments I wish to thank Dr. P. Kaesberg, University of Wisconsin (Madison), for permission to use the BBV sequence in this study. I also want to express my appreciation to A. Hart for technical assistance.
t9 y . N. Chang, D. L. Y. Dong, G. S. Hayward, and D. Hayward, J. Virol. 64, 3358 (1990). 2o A. J. Shatkin, Cell 9, 645 (1976). 2r M. Kozak, Nature (London) 3118, 241 (1984). z2 p. D. Friesen and R. R. Rueckert, J. Virol. 49, 116 (1984). 23 p. D. Friesen, P. Scotti, J. Longworth, and R. R. Rueckert, J. Virol. 35, 741 (1980).
152
VECTORS FOR EXPRESSING CLONED GENES
[1 1l
[11] Cell-Free Expression of Large Collections of cDNA Clones Using Positive-Selection h Phage Vectors By
CHRISTOPHER COLECLOUGH
Cell-free transcription and translation is frequently used to gain information about the structure and activity of proteins encoded by cDNA clones. Hitherto it has generally been applied to specific, individual clones, identified and isolated by some other means. The analogous procedure of microinjection of translationally competent RNA transcribed from collections of recombinants has, however, been used successfully in isolating clones encoding proteins with known biological activities (see Ref. 1 for an example). Here I describe vectors and technology appropriate to the analysis of large groups of clones as well as individual clones, in terms of their protein-coding capacity. This methodology can be used not only in searches for clones encoding specific proteins, but also in the analysis of complex mRNA populations. This chapter is concerned chiefly with technical details, but some applications of the methodology will be discussed. Vector Design A general problem in cDNA cloning is the dilution of true recombinants in libraries by nonrecombinant genome types, lacking cDNA sequences. At present, most libraries are made with a view to recovering one or a few particular clones for which specific probes--nucleic acid or antibody--are available. As the plaque-screening methods used can usually be applied to very large numbers of plaques, a high frequency of nonrecombinants is therefore most often accommodated by simply making larger libraries. A low incidence of true recombinants, however, becomes a severe problem if more complicated screening schemes are envisaged. The frequency of nonrecombinant types can be greatly reduced if true recombinants develop a selectable phenotypic trait not possessed by the parental vector. ~gtl0 is the most commonly used vector that allows such selection. 2 It is a temperate phage, which forms turbid plaques on its usual host Escherichia coli strain, due to partial repression. On hfl (highi y . Noma, P. Sideras, T. Naito, S. Bergstedt-Lundquist, C. Azuma, E. Severinson, T. Tanabe, T. Kinashi, F. Matsuda, Y. Yaoita, and T. Honjo, Nature (London) 324, 70 (1986). 2 T. V. Huynh, R. A. Young, and R. W. Davis, in "DNA Cloning: A Practical Approach" (D. M. Glover, ed.), Vol. I, p. 49. IRL Press, Oxford, England, 1985.
METHODS IN ENZYMOLOGY,VOL. 217
Copyright© 1993by AcademicPress, Inc. All rights-ofreproductionin any formreserved.
[11]
h PHAGEVECTORS
153
frequency lysogeny) varients ofE. coli, however, repression is so efficient that plaque formation by parental hgtl0 is entirely suppressed, cDNA insertion into the hgtl0 genome is directed to the cI gene (which encodes the phage repressor) and when this gene is thus disrupted, the inability of recombinants to elaborate active repressor leads to clear plaque formation, even on hfl strains. Although this device is quite effective, it places an absolute restriction on the context into which cDNA can be inserted. It therefore precludes, for example, juxtaposition of cDNA with a promoter for T7 or SP6 RNA polymerase, desirable if in vitro expression is planned. (Modified versions of hgtl0 that include these promoters do exist, but they have lost the capacity for genetic selection.) I have constructed a family of vectors that allows a strong genetic selection for true recombinants, yet is much more flexible with regard to the context into which cDNA can be inserted. This is because the phenotypic difference between parental and recombinant types results in part from the acquisition of a cDNA-linked marker, rather than from the disruption of a vector function. This marker is the chi recombination target, which being only 8 nucleotides long can easily be incorporated into the oligo(dT)-containing primer used in cDNA synthesis. Two vectors, hjac and hecc, are illustrated in Fig. 1; they are based on hgtWES.hB 3 and inherit its general structure and amber mutations, differing from it around the site of cDNA insertion. These vectors are r e d - but contain functional gam genes, so are spi ÷ (sensitive to P2 interference) and do not form plaques on E. coli lysogenic for phage P2. Preparation of hjac and hecc to receive cDNA inserts deletes the g a m gene; r e d - gam - h phage are s p i and will form plaques on P2 lysogens, but grow exceedingly poorly (when host recombination systems are active) unless they contain chi sites, which hjac and hecc lack. The necessary chi sites are provided in the primer-restriction end adapter (PRE adapter) used to initiate cDNA synthesis. (See Stahi 4 for a review of spi and chi.)
Primer-Restriction End Adapters Several years ago, to reduce the number of steps in cDNA cloning schemes, we introduced the use of bifunctional PRE adapters that serve both as primers for reverse transcription and as restriction ends for ligation to vectors. 5'6 The PRE adapters used to insert cDNA into hjac and hecc 3 p. 4 F. C. 6 C.
Leder, D. Tiemeier, and L. Enquist, Science 196, 175 (1977). W. Stahl, Sci. Am. 2S6, 53 (1987). Coleclough and F. Erlitz, Gene 34, 305 (1985). Coleclough, this series, Vol. 154, p. 64.
154
[11]
VECTORS FOR EXPRESSING CLONED GENES
cosL]
nin5 Wam
Eam
~gtWES.,
35
36
PL
;P6
Xjac XZ ,ST SZ
SZ NZ~ t
I
i
TSP6~PT7 PL
. . _ T T 7 PSP6~ (zm
t
~,
PL
i~
,1.ecc NZ s'r
SZ
SZ XI"
FIG. 1. Structure of hjac and hecc. hjac and hecc are based on hgtWES.hB, the structure of which is illustrated, highlighting the areas altered in the new vectors. Numbers 32-36 indicate the sequence coordinates on the conventional h map, not the physical distance from the left-hand end. PL, h phage PL promoter; PSP6, promoter for SP6 RNA polymerase; TSP6, terminator for SP6 RNA polymerase; PT7, promoter for T7 RNA polymerase; TT7, terminator for T7 RNA polymerase. Restriction enzyme cleavage sites: ERI, EcoRI; NI, NotI; SI, Sail; XI, Xhol.
are shown in Fig. 2. The adapters are used in two ways: (1) to convert a staggered restriction end into a 3' homopolymer tail, and (2) to provide cDNA molecules both with a chi site and with a staggered restriction end suitable for ligating to vector molecules. Each function requires a pair of partially complementary partner oligonucleotides. PRE adapters can be designed for sequence-specific or, as here, general cDNA synthesis (see Ref. 6) and for ligation to any staggered restriction end. Oligo(dC) tailing of ~,jac requires a pair of PRE adapters, (iii) and (iv) in Fig. 2, suitable for ligation to XhoI ends, while a pair compatible with NotI ends, (v) and (vi) in Fig. 2, is used for hecc. Pair (i) and (ii) in Fig. 2, used for cDNA synthesis and ligation to a SalI end, can be used with either vector. Following standard solid-phase synthesis, elution, and deprotection of the oligonucleotides, purify them by polyacrylamide gel electrophoresis (PAGE), DEAE-cellulose chromatography, and filtration on Sephadex G-50 superfine. Adapters (i), (iii), and (v) should be fully 5' phosphorylated before use. Incubate them at 1-5/zM with 50-100 units of T4 polynucleotide kinase in l mM ATP, 50 mM Tris-HC1 (pH 7.6), 10 mM MgCI2, 5 mM dithiothreitol (DTT), 1 mM spermidine, 1 mM ethylenediaminetetraacetic
[1 II
h PHAGE VECTORS
155
(i) (ii)
ch£ 5' TCGAC~CCACCAGCTCTTTTTTTTTTTTTTTT 3' GTtGGTGGTC~AGAAAAA 5'
(iii) (iv)
5' TCGAGTCTAGACGCGTTCCCCCCCCCC 3' CAGATCTGCGC 5'
(v) (vi)
5' GGCCGCTCTAGATCTCTTCCCCCCCCCC 3' CGAGATCTAGAG 5'
3'
3'
3'
FIG. 2. PRE adapters for use with kjac and hecc. Set (i) and (ii) provides a primer for reverse transcription, a SalI end, and contain a chi site. Set (iii) and (iv) converts a Xhol end into an oligo(dC) tail and is used for tailing hjac. Set (v) and (vi) converts an Nod end into an oligo(dC) tail and is used with hecc.
acid (EDTA), for 3 hr at 37 °. R e c o v e r the oligonucleotide by filtration through G-50 superfine. T o check that phosphorylation is efficient, set up a parallel reaction with a small quantity of the starting oligonucleotide and a known molar excess of ATP to which is added some fresh [y-32p]ATP: the incorporation of radiolabel should approximate the molar ratio. Store the oligonucleotides in 1 m M E D T A at - 2 0 °.
Vector Preparation Preparation of hjac or hecc D N A to receive e D N A inserts consists of four steps, illustrated in Fig. 3, which for simplicity shows only the preparation of hjac; the preparation of hecc is similar, but uses different restriction e n z y m e s and adapters, as will be described. The steps are the following: (I) restriction e n z y m e digestion, (2) ligation to PRE adapters, (3) digestion with a second restriction enzyme, and (4) removal of small fragments. 1. Digest kjac D N A with X h o I , hecc D N A with N o t I , to completion; r e c o v e r D N A by p h e n o l - C H C l , extraction, and ethanol precipitation. 2. For e v e r y 100/xg of digested, redissolved vector DNA: add 3/xg of adapter (iii) or (v) and 1.3/xg of adapter (iv) or (vi)--using (iii) and (iv) for hjac; (v) and (vi) for h e c c w r o u g h l y a 40-fold molar excess of PRE adapters over vector ends. Treat with T4 D N A ligase at 150 Weiss units/ml in 50 m M Tris-HCl (pH 7.8), 10 m M MgCI:, 10 m M DTT, 1 m M ATP, 100/xg/ ml bovine serum albumin (BSA) for 16 hr at 16° in a final volume of I00 /xl. Precipitate vector DNA, now ligated at cos ends and with two oligo(dC)
156
[11]
VECTORS FOR EXPRESSING CLONED GENES
~.jac: digest with XhoZ anneal and Ilgate to (il) and (iv) 11am
1
cos
cccccl.II II
mRNA: reverse transcribe using (i) as primer. Tail with TdTase
ccccc
and d GTP
SalZ
1
digest with SalZ, remove small fragments
3" GGGG~%,.-~J-~TTT I~JAGC~ 5"
!
TCGA
-C06 ~--
fractlonate, if desired CCCCC
J anneal with (ll) Ilgate and gap-fill
1
COS
TCGA" - ~
CCCCC
AA~
TCGA~
COS
CCCCC
package in vitro
1 plate on I:>2lysogen FIG. 3. Flow chart for inserting cDNA into hjac.
tails, by addition of NaCI to 0.5 M and polyethylene glycol (PEG) 6000 to 8%, 7 3. Digest DNA with SalI to completion. 4. Separate the finished vector (39 kb) from the smaller digestion products [0.7, 0.5 kilobases (kb)], by centrifugation through a 10-40% sucrose gradient. 8 mRNA Isolation A library intended for cell-flee expression should ideally be constructed from a pure mRNA preparation. In practice, it is impossible to state with confidence that any RNA preparation contains only mRNA, that is, only those molecules actually translatable into complete polypeptides. A rigorous mRNA preparation protocol would include careful purification of polysomes free from other ribonucleoprotein (RNP) particles, specific release and purification of mRNP from the polysomes, and purifi7 j. Lis, this series, Vol. 65, p. 347. 8 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989.
[11]
h PHAGEVECTORS
157
cation of mRNA from mRNP, avoiding throughout any nucleolytic cleavage. In contrast, it has become standard practice to recover " m R N A " from guanidinium lysates of cells, neglecting any subcellular fractionation. The following is a compromise that removes much, but not all, unwanted RNA of nuclear origin and is appropriate for all but the most nucleaseridden sources. Cells are lysed with nonionic detergent, in the presence of vandium-ribonucleoside complexes to inhibit RNase, 9 and nuclei are pelleted before extracting RNA from the supernatant. Artifactual degradation can be further avoided by keeping samples cold and working quickly. Use of a microfuge greatly speeds up the process so that the initial steps take only 2-3 min, and should be used for l0 s cells or less; if the microfuge is at room temperature, remove and prechill the rotor head. Vanadylribonucleoside complex (VRC) should be prepared, not purchased, carefully following published protocols for avoiding oxidation, J0 and stored in aliquots under liquid N 2 . Wash cells in 0.9% (w/v) NaC1. Using about 1 ml/10 s cells, resuspend cells in ice-cold 10 mM Tris-HCl (pH 8.6), 12 mM MgCI 2, 10 mM VRC. Add 20-50/~l/ml cold 10% (v/v) Triton X-100 and vortex 5 sec, or homogenize using a Dounce homogenizer for 6 strokes. Spin at 10,000 g for 5 rain in a preparative centrifuge, or 15 sec in a microfuge. Transfer supernatant and add, for each milliliter, 30/zl of 5 M NaC1, 50/zl of 0.5 M EDTA, 50 kd of 20% technical grade sodium dodecyl sulfate (SDS) and 0.6 ml of phenol/CHCl3/8-hydroxyquinoline (50 : 50 : 0.2), saturated with 0.1 M NaC1, 10 mM sodium acetate, pH 6.0, 1 mM EDTA.~I Mix vigorously for 5 sec and spin at room temperature to separate phases. Reextract the upper, aqueous phase after addition of 20/.d/ml 20% SDS by shaking 2 min with an equal volume of phenol/CHC13 . Repeat the extraction, with 3 min of shaking, and precipitate RNA by addition of 2 vol of ethanol and storage at - 2 0 °. This lysis buffer itself will throw a bulky ethanol precipitate, probably a mixture of sulfates. This does not interfere with oligo(dT)-cellulose chromatography; however, if it is desired to avoid the chromatography step, it may be necessary to purify the RNA further, or to omit VRC from the lysis buffer. Purify poly(A)-containing RNA by passage over oligo(dT)-cellulose. It is important to find a grade of highcapacity oligo(dT)-cellulose that will efficiently retain mRNA in 0.15 M NaC1, or less. Use a small column: a 0.5-ml bed should be sufficient to recover mRNA from up to 20 mg of cytoplasmic RNA. After two cycles 9 S. L, Berger and C. S. Birkenmeier, Biochemistry 18, 5134 (1979). ~0G. E. Lienhard, I. I. Secemski, K. A. Koehler, and R. N. Lindquist, Cold Spring Harbor Symp. Quant. Biol. 36, 45 (1971). fl R. P, Perry, J. La Torre, D. E. Kelley, and J. R. Greenberg, Biochim. Biophys. Acta 262, 220 (1972).
158
VECTORS FOR EXPRESSING CLONED GENES
[1 II
of binding and elution, separated by a heat treatment of 2 min at 65 °, remove SDS and any fines by adding potassium acetate to 0.2 M and shaking with phenol/CHCl3 ; precipitate RNA by adding 2 vol of ethanol. Store purified mRNA at 1 mg/ml in diethyl pyrocarbonate-treated water, in aliquots at - 7 0 °. cDNA Preparation More details of this reaction can be found in the article in this series by Coleclough. 6 The quality of the reverse transcriptase is the principal determinant of the quality of a cDNA library. Use XL grade AMV reverse transcriptase from Life Sciences (St. Petersburg, FL). It is convenient to dispense the enzyme into 5-/zl aliquots of 65 unitsmsufficient enzyme to copy 5 /zg of mRNA--diluting the reverse transcriptase with RNasin (Promega, Madison, WI), if necessary; store them at - 7 0 °. Prepare 10 × RT buffer by combining 0.7 M KCI, 0.5 M Tris-HCl, pH 8.78 at 25 °, and 0.1 M MgC1z ; store at - 20° in aliquots. To 5/zl of 1 mg/ml mRNA, add 5/zl of freshly diluted 20 mM methylmercuric hydroxide (Alfa, Ward Hill, MA). After 2 min at room temperature, transfer the reaction tube to an ice bath and add, in this order: 10/xl of a 20-/xg/ml solution of PRE adapter (i); 5/~1 of a solution of dATP, dCTP, dGTP, and dTTP (20 mM each); 1/~1 of fresh [a-3ep]dCTP (I0 mCi/ml); 1 /zl of water; 5 /xl of 10 × RT buffer; 6 ~1 of freshly diluted 300 mM 2-mercaptoethanol; 2.5 /zl of actinomycin D (1 mg/ml); 0.5 ~1 of polynucleotide kinase (10,000 units/ml); 4 tzl of RNasin (I0,000 units/ml); and 5/~1 of reverse transcriptase, aliquoted as above. Incubate 2 min at 16°, 20 min at 43 °, and 5 min at 48 °. Add 2/zl of 0.5 M EDTA, 3 tzl of 20% (w/v) SDS, 1 tzl of proteinase K (10 mg/ml), and incubate 20 min at 48 °. Add 1/xl 100 mM phenylmethylsulfonyl fluoride (PMSF) in dimethyl sulfoxide (DMSO), chill, extract with phenol/CHCl3, and ethanol precipitate. Dissolve the pellet in 20/xl of 0.2 M NaOH and 1 mM EDTA, and incubate at 60° for 20 min. Cool and dilute the sample with water to 50/xl, then load onto a calibrated column of Sephacryl S-500 HR (Pharmacia, Piscataway, NJ), previously washed in 10 mM NaOH. Collect and pool cDNA longer than 300 nucleotides. Neutralize and concentrate the solution to 50 txl by extraction with 2-butanol, then recover cDNA by ethanol precipitation. The object of the gel-filtration step is to eliminate PRE adapter molecules not incorporated into cDNA. S-500 columns formed in siliconized Pasteur pipettes can resolve single-stranded DNA chains a few hundred nucleotides long well enough to achieve this and, with appropriate care, are sufficiently reproducible that calibration of one of a batch of columns
[11]
X PHAGEVECTORS
159
with marker DNA fragments should allow removal of low molecular weight species, without the need for analysis of all fractions. The yield of cDNA can be calculated from the molar incorporation of dCTP and should be 2-3 tzg. For reactions of this size, it is worth checking the yield spectrophotometrically, as the incorporation of nucleotide may be misleadingly low. The recipe given above generates cDNA with a specific activity of (very roughly) 30,000 Cerenkov counts per minute (cpm)//xg. cDNA is elongated with a 3'-oligo(dG) tail by treatment with terminal transferase: adjust volumes so that the final cDNA concentration is 60 tzg/ ml (or less) in 50 # M dGTP, 600 units/ml terminal transferase, and 100 mM potassium cacodylate; 25 mM Tris base; 1 mM CoC1 z; 0.2 mM DTT, pH about 6.9 (see Ref. 12); incubate exactly 3 min at 16°, then add EDTA to 10 mM, SDS to 1% (w/v), extract with phenol/CHC13 and ethanol precipitate; redissolve the cDNA in 10 mM NaOH and I mM EDTA, filter it through Sephacryl S-500 HR in 10 mM NaOH, neutralize, concentrate, and ethanol precipitate it, all as above. Single-stranded DNA at low concentration tends to stick to surfaces and interfaces, so whenever possible these manipulations should be performed on cDNA in quantities of 1 /zg or more; parceling out a cDNA preparation into samples of about 100 ng, with a view to avoid wasting it, is more likely to result in the loss of the entire sample. cDNA Fractionation by Subtractive Hybridization Insertion of cDNA into hjac or hecc is performed in approximate molar equivalence of cDNA and vector. As this requires about a 70-fold mass excess of vector DNA, use of much more than 100 ng of cDNA becomes prohibitively expensive. Therefore, as more cDNA will probably be made than can be used, unless the application demands construction of a representative library, it is worthwhile considering ffactionating the cDNA, by size or by molecular hybridization, before cloning. We have found hybridization with biotinylated mRNA, followed by liganding with streptavidin and phenol/chloroform removal of complexes, ~3 to be a simple, efficient, and high-yield tactic for enrichment of differentially expressed sequences prior to cloning. mRNA is biotinylated by treatment with a photoactivatable biotin derivative, as described by Forster e t al. 14 RNA at l mg/ml in water is ~2R. Roychudhury and R. Wu, this series, Vol. 65, p. 43. 13H. L. Sive and T. StJohn, Nucleic Acids Res. 16, 10937(1988). 14A. C. Forster, J. L. Mclnnes, D. C. Skingle, and R. H. Symons, Nucleic Acids 745 (1985).
Res.
13,
160
VECTORS FOR EXPRESSING CLONED GENES
[11]
mixed under subdued light with an equal volume of photobiotin (Calbiochem, La Jolla, CA) dissolved at 1.5 mg/ml in water; the mixture is sealed in a siliconized glass capillary and exposed to light from a 275-W sunlamp bulb (available from Bethesda Research Laboratories, Gaithersburg, MD) 10 cm away for 20 min. The solution is kept cold by immersion in a shallow ice/water bath. Biotinylated RNA is recovered by flushing out the contents of the capillary with 100/zl of 0.1 M Tris-HCl, pH 9.0, twice extracting with n- or 2-butanol, and precipitation with 2 vol of ethanol after addition of NaCI to 0.2 M. Hybridization reactions contain biotinylated mRNA at 0.5-2.0 mg/ml, 250/xg/ml oligo(C), 2 mg/ml poly(A), 0.6 M NaCI, 50 mM piperazineN,N'-bis(2-ethanesulfonic acid) (PIPES) (pH 6.8), 5 mM EDTA, and 0.5% (w/v) SDS and are held at 65° typically for 5-15 hr. Optimally, such reactions are constantly agitated, for example, by rotating the mixture sealed in a capillary tube with a few glass beads, but they can be performed in tightly capped microcentrifuge tubes with some loss of efficiency. The efficiency of RNA-driven hybridization of DNA and the effects of salt concentration and other parameters on the hybridization rate are fully discussed by Van Ness and Hahn. ~5 Oligo(C) is prepared by limited hydrolysis ofpoly(C)~6:4 mg ofpoly(C) in 10 mM Tris-HC1, pH 7.5, is mixed with 0.5 ml of 0.4 M NH4HCO3/ NH4OH, pH I0.0, and the mixture sealed in a glass ampoule and placed in a boiling water bath for 1 hr. At the completion of the hybridization reaction, the mixture is diluted 30-fold with water, BSA is added to 300/zg/ml, and streptavidin to 30 /zg/ml. The mixture is then vortexed with an equal volume of phenol/ chloroform and spun in a microcentrifuge for ! min. After reextraction with 1 vol of phenol/chloroform, nucleic acids are precipitated from the aqueous phase with 2.5 vol of ethanol. We have usually used two rounds of subtractive hybridization before cloning. Finally, cDNA is freed of RNA by alkaline hydrolysis and separation on Sephacryl S-500 HR, as above. Insertion of cDNA into Vector Molecules cDNA insertion is a one-tube, two-stage reaction (see Fig. 3). In the first stage, the 5' end ofcDNA is rendered partially duplex by hybridization of the PRE adapter (i) sequence, which forms the cDNA 5' terminus, with 15 j. Van Ness and W. E. Hahn, Nucleic Acids Res. 10, 8061 (1982). 16j. M, Coffin, J. T. Parsons, L. Ryrno, R. K. Haroz, and C. Weissman, J. Mol. Biol. 86, 373 (1974).
[11]
,k PHAGE VECTORS
161
the partner adapter (ii), then covalently attached to the SalI end of the vector. At the same time, the 3' oligo(dG) tail of the cDNA hybridizes to the other, oligo(dC) vector terminus, cDNA, thus forming a singlestranded bridge linking left-hand and right-hand X arms. In the second stage, cDNA is at last rendered double stranded in a gap-filling reaction of DNA polymerase, and remaining nicks are sealed. Figure 4 illustrates the structure of a typical hjac recombinant, which was recovered from a small library of about 5000 clones made from 2 ng of cDNA. This clone encodes the T cell receptor/3 chain expressed by the AKR mouse thymorea line BW5147, frequently used as a T cell hybridoma parent. A representative reaction was performed as follows: 60 ng of tailed cDNA and 4.5/xg of appropriately prepared hecc DNA were combined in 20 txl together with 5 ng of adapter (ii) and 2.5/xl of 1 M Tris-HCl, pH 7.8, and 100 mM MgC12. This mix was annealed by placing the reaction tube in a beaker of water, initially at 60°, which was allowed to cool to 20° over the course of about 1 hr. The reaction volume was then increased to 25 ~zl with the addition of BSA to 50/xg/ml, DTT to 10 raM, ATP to 300/xM, and 3 Weiss units of T4 DNA ligase, and incubated at 16°C for 16 hr. The second stage of the reaction was performed in conditions of molecular crowding by polyethylene glycol (PEG)iV: to the 25-/xl first-stage reaction was added 15/xl of 50% PEG 6000 and 5/zl of 10 x stage 2 buffer (see below), and the volume was brought to 50 ~!, adding dATP, dCTP, dGTP, and dTTP (each to 500 ~M), 3 units each ofE. coli DNA polymerase I and T4 DNA polymerase, and 0.2/xg of E. coli DNA ligase. To prepare 10 x stage 2 buffer, combine 1 M NaCI, 0.02 M (NH4)2SO4, 0.1 M Tris-HCl (pH 7.5), 60 mM MgC12,50 mM DTT, 10 mM spermidine, and 3 mM NAD. The reaction was allowed to stand for 3 hr at room temperature, then DNA was pelleted by spinning in a microfuge for 2 rain and resuspended in 4 /xl of water. Ligated DNA can now be packaged into infectious )t particles using a commercial packaging kit. Unfortunately, the in vitro packaging reaction, treated here as a black box, is the most sensitive step in the whole process: when cloning attempts yield greatly fewer plaques than expected, this is most likely due to inhibition of the packaging reaction. DNA purified from agarose gels can carry contaminants--probably charged polysaccharides--which do not affect any of the enzymes used, but profoundly inhibit packaging. Also, carryover of PEG into the packaging reaction must be minimized: at low concentrations PEG is noninhibitory, but slight increases can precipitate components of the packaging mix with disastrous results. The cDNA-vector i7 S. B. Zimmerman and B. Harrison, Proc. Natl. Acad, Sci. U.S.A. 84, 1871 (1987).
162
VECTORS FOR EXPRESSING CLONED GENES
zl
CO
~
o
0 0 0
0 0
,-.= "-~
,]i °
°° Ig
oo G.~
X
0 0 < Z
£ 0
~._o
[ l 1]
[11]
h PHAGEVECTORS
163
ligation reactions should therefore be performed in microfuge tubes that resist wetting, allowing complete removal of the supernatant after pelleting the ligated DNA. Forming a Sectored Library The selective host for spi- recombinants of hjac and hecc is LE392/ P2, a P2 lysogenic derivative of the supE supF E. coli strain LE392. Grow these bacteria in L broth [1% (w/v) tryptone, 0.5% (w/v) yeast extract, 0.5% (w/v) NaCI] containing 0.4% (w/v) maltose and 10 mM MgSO4 to stationary phase at room temperature, shaking vigorously, then pellet them at 2000 g for 10 min and resuspend them in a two-thirds volume of 10 mM MgSO4. Dilute packaged, recombinant phage to 0.5 ml with phage buffer and plate no more than 20 ~1 of this suspension on one 100-mm petri dish, using 0.2 ml of LE392/P2 and 2.5 ml of 0.6% (w/v) top agar. Mix the phage and bacteria and incubate for 20 min at 37°, then add the agar (cooled to 43 °) and plate. Phage buffer contains 0.1 M NaCI, 25 mM Tris-HCl (pH 7.5), 10 mM MgSO4, 0.02% (w/v) gelatin. Bottom agar for plates is 1.2% (w/v) agar in L broth containing 10 mM MgSO4 and 0.1% (w/v) maltose; top agar contains 0.6 or 0.3% (w/v) agar, as indicated, in the same medium. The quality of the packaging mix is the primary determinant of the efficiency of clone formation; about 2000 clones/ng of cDNA is routine. A sectored library is desirable for many applications using these vectors. Division of the library into sectors is preferably performed on the primary plates. Sectors can be of any complexity; 500-1000 clones per sector is probably the most generally useful size for nonenriched libraries. Decide the number of pie-type sectors into which each plate should be divided, draw a template on the lid of a petri dish, and use this as a guide for each plate. Using a straight-edged nickel spatula, scrape the top agar from each sector into a 15-ml polypropylene tube, taking care not to touch other sectors. Between each sector, wash the spatula in 80% ethanol, flame it, and rinse it in 10 mM MgSO4. To the collected agar in each tube add 1 ml of phage buffer containing 20% (v/v) glycerol and 1 ml of CHCI 3 ; cap and shake it for 20 rain. Spin the tubes at 2000 g for 10 rain, transfer the clear supernatant to vials, and store them at - 70°. This is the primary sectored library. Expanding Sectors and Clones Single clones, sets of clones, and library sectors are all expanded in the same fashion to produce a standard, high-titer stock that can be used
164
VECTORS FOR EXPRESSING CLONED GENES
[11]
to seed cultures for DNA preparation. To 150/~1 of a suspension of LE392/ P2, prepared as above, add the equivalent of one plaque of phage and after absorption plate with 3 ml of 0.3% (w/v) top agar. Use a fairly fresh plate. A bacterial lawn should be visible after 4 hr, and confluent lysis evident after about 7 hr. Scrape the sloppy top agar into a 15-ml polypropylene tube, add 1 ml of phage buffer containing 50% (v/v) glycerol and 2-3 ml of CHC13, and shake and spin as above. This generates a phage stock of about 101~plaque-forming units (pfu)/ml, stable at - 70 °; expanded primary library sectors are termed secondary sectors. Preparing DNA for Cell-Free Expression Use LE392/P2 for all growth of spi- recombinants on plates, and LE392 for growth in suspension culture. Reportedly spi-h grows better on recD E. coli, so substitution of LE392 by the supE supF recD strain TAP90 ~s might well increase the phage yield from these cultures; I have not compared the two strains. Prepare LE392 for infection as above. Allow 1 /~1 of high-titer phage stock to absorb to 250/xl of bacteria in a 50-ml polypropylene tube. Add 20 ml of L broth containing 10 mM MgSO 4 and 0.01% (w/v) maltose and shake the tube vigorously in a horizontal position at 37°. Lysis should be evident after 5 hr. Add a drop of CHCI 3, spin out debris (there may be very little), and precipitate phage by addition of 1.2 g of NaC1 and 2 g of PEG 6000. Hold at least 1 hr on ice. Pellet phage at 3000 g, for 10 rain at 4°, and resuspend in 800/zl of 10 mM Tris-HCl, pH 7.5 and 10 mM MgCI2 containing 10/.tg/ml each of DNase I and RNase A. Shake at 37° for 10 min. Add 0.5 ml of CHCI3, mix well (do not vortex), spin briefly, and transfer the supernatant to a microfuge tube. Add 25/zl of 0.5 M EDTA, 50/xl of 20% (w/v) SDS, and 12/zl of proteinase K (10 mg/ml), and incubate 45 min at 37°. Add I0 gl of I00 mM PMSF in DMSO, hold 5 min, then add 50 /zl of 4 M KC1, extract with phenol/CHCl3, shaking by hand, and precipitate DNA with 0.6 ml of 2-propanol. The routine yield is 7-10/zg of DNA, which is fine for restriction enzyme analysis and, often, for transcription too; the degree of contamination with degraded bacterial nucleic acid, however, may be sufficient strong to inhibit RNA polymerase. It would probably be worthwhile investigating other, more selective agents for precipitating large DNA, such as PEG 7 or spermine, 19 with a view to eliminating this contamination, but gel filtration on Sephacryl S-1000, which is effective if somewhat tedious, has been routinely in~s T. A. Patterson and M. Dean, Nucleic Acids Res. 15, 6298 (1987). 19 B. C. Hoopes and W. R. McClure, Nucleic Acids Res. 9, 5493 (1981).
[11]
X PHAGEVECTORS
165
cluded, k DNA reproducibly begins to elute after 0.9 ml of washing of a Pasteur pipette S-I000 column, while small nucleic acid fragments appear after passage of 1.5 ml. This fraction can be collected from many columns run simultaneously in 30 mM potassium acetate, 5 mM Tris-HC1, pH 7.5, and 0.5 mM EDTA, and concentrated about fivefold by 2-butanol extraction; then the DNA may be recovered by ethanol precipitation and dissolved at about 100/xg/ml in 10 mM Tris-HC1, pH 7.5, 1 mM EDTA.
Transcription Sense-strand, translatable transcripts of recombinants of hjac or hecc are referred to as "ersatz m R N A " to emphasize that they can substitute for genuine, natural mRNA for most purposes. Ersatz mRNA is generated from hjac recombinants with SP6 RNA polymerase, and from hecc recombinants with T7 R N A polymerase. Both vectors have fairly effective terminators immediately distal to the site of cDNA insertion, so it is not necessary to digest DNA with restriction enzymes to produce discrete transcripts. A standard transcription reaction is 30/xl, which includes 6/zl of DNA, prepared as above, 500/xM ATP, CTP and UTP, 150/xM GTP, and 500/xM m7GpppG. It is convenient to add about 1/zCi of [a-3Zp]GTP to each reaction to help trace the RNA product. Final reaction conditions for SP6 RNA polymerase are as follows: 40 mMTris-HCl (pH 7.9); 6 mM MgCI2, 10 mM DTT, 100/xg/ml BSA, 500 units/ml RNasin. Conditions for T7 RNA polymerase are similar, except that the MgCI2 concentration is 10 raM. Use 1/zl of either polymerase, as obtained commercially, for each reaction; this usually contains about 5 units of the SP6 enzyme, but much more T7 polymerase. The T7 enzyme is produced commercially from synthetic constructs in E. coli, and at the moment is routinely supplied at much higher activity than the SP6 enzyme. Reactions are incubated at 37° for 90 rain. Ersatz mRNA is most simply purified by binding to a tiny oligo(dT)-ceUulose column: to the 30/zl transcription reaction, add 1/xl of 0.5 M EDTA, 2/xl of 20% (w/v) SDS, 5/.tl of 5 M NaCI, and 12 of/xl water, and extract with 50/xl of phenol/CHCl s. Apply the extract directly to a small (about 20 p.l) column of oligo(dT)-cellulose, formed in a plugged 1-ml micropipette tip. Wash the column with 200/zl of 0.5 M NaCI, 10 mM TrisHCI (pH 8.0), 1 mM EDTA, 0.2% (w/v) SDS, and elute with applications of 15, 15, and 70/xl of 5 mM Tris-HCl (pH 8.0), 0.5 mM EDTA, 0.1% (w/v) SDS, using the first two batches to rinse the sides of the column and pooling the eluates. Add I/xg of calf liver tRNA (Boehringer Mannheim, Indianapolis, IN) and 5/xl of 5 M potassium acetate to the eluted RNA, extract with 100/xl of phenol/CHC13 and precipitate with 2 vol of ethanol.
166
VECTORS F O R E X P R E S S I N G C L O N E D G E N E S
[11]
Typical yields are 0.5-1.0/.,g of ersatz mRNA when SP6 polymerase is used, and 2-3/zg with the higher-activity T7 enzyme. The cap dinucleotide can be omitted for RNA to be used for some other purpose than translation. Even though all transcripts from a ~,ecc, or from a hjac, library share a common 5' terminal sequence--contributed by the vector and by the PRE adapter used for tailing the vectormdifferent clones vary greatly in their dependence on a 5' cap for translation. Some are unaffected by the lack of a cap, while the translation of others is entirely cap dependent; therefore, unless the object is to assay a species known to be cap independent, all ersatz mRNA for translation should be transcribed in the presence of m7GpppG. Translation For routine analytical purposes, rabbit reticulocyte, nuclease-treated lysate #N90 from Amersham (Arlington Heights, IL) is used to translate ersatz and genuine mRNA, using [35S]methionine as label. A typical reaction will include all of the product of a standard SP6 transcription reaction, or one-third of a T7 transcription, in a final volume of 20 ~1, of which 16 /zl is reticulocyte lysate. Incubate for 90 min at 30°, then add 1 /.d of RNase A (2 mg/ml; to degrade aminoacyl-tRNA), and incubate for 10 min. Polypeptide products can now be displayed on polyacrylamide gels; it must be borne in mind that the hemoglobin concentration in the translation reaction is about 60 mg/ml, limiting the fraction of the reaction that can be applied to a typical gel lane. The routine here is to dilute 4/zl with 25 /zl of 3% (w/v) SDS loading buffer for application to a 4 × 0.75 mm well. The reticulocyte lysate will generate a labeled protein complex endogenously, which usually has an apparent molecular weight of about 45K on SDS-PAGE, when methionine is used as label. [35S]Cysteine does not label this complex. Two-dimensional (2D) gel patterns of in vitro translation products of the same mRNA preparation labeled alternatively in methionine or in cysteine are surprisingly different, those labeled with [35S]cysteine usually showing the greater number of spots; this suggests that methionine and cysteine should both be labeled when complete representation of the complexity of an mRNA sample is desired. Discussion Ersatz mRNA, transcribed from cDNA clones, can substitute genuine mRNA in any application; thus it can be translated, hybridized, reverse transcribed, and microinjected. A vast array of experimental possibilities is therefore available for the exploitation of libraries in hjac or hecc;
[11]
x PHAGEVECTORS
167
here examples are shown only of the use of gel electrophoresis to probe relatively complex collections of clones. The analysis of proteins, translated in vitro from natural or ersatz mRNA, can be viewed in two ways: (1) in the absence of a high-resolution separation method for large R N A molecules, complex collections of mRNA molecules are best analyzed by translation into proteins for which high-resolution analytical methods do exist, and (2) in vitro translation products may resemble naturally biosynthesized proteins sufficiently closely as to allow the identification of cDNA clones that encode proteins of interest through the physical or biological properties of the cell-free translation products of ersatz mRNA. The first of these statements, which does not presuppose any identifiable relationship between in vivo and in vitro translation products of the same mRNA species, is certainly true. The second, and more exciting, is true in some but not all cases. As yet we do not know to what fraction of cases it applies, although we know it is a sizable fraction. It may well turn out to be possible, with adequate attention, to relate the great majority of proteins physically detectable in cell extracts directly to cognate cDNA clones, which can be retrieved without the need for sequence information or sequence-specific probes. 2°'2~
Two-Dimensional Gel Analysis of Translation Products Running many 2D gels with high and reproducible resolution demands considerable expertise. The examples shown here were run in the laboratory of Dr. I. Lefkovits, of the Basel Institute for Immunology. Technical details for 2D gel electrophoresis or computer-assisted image analysis have not been included in this chapter; both of these aspects have generated a considerable technical literature, which should be consulted by potential users. However, occasional 2D gel electrophoresis should be within the compass of most laboratories; it is most easily accomplished in commercially available equipment following the recommendations of the supplier, and it is extremely valuable for the analysis of hecc or Xjac libraries, even if the quality of the gels is less than pristine. This is particularly so in evaluating "difference" libraries, formed from cDNA that has been enriched for species of interest by subtractive hybridization. The construction of difference libraries, enriched for some cDNA species of interest, has become a popular tactic, especially since its successful 20 C. Coleclough, L. Kuhn, and I. Lefkovits, Proc. Natl. Acad. Sci. U.S.A. 87, 1753 (1990). 21 I. Lefkovits, J. Kettrnan, and C. Coleclough, lmmunol. Today 11, 157 (1990).
168
VECTORS FOR EXPRESSING CLONED GENES
[11]
use in the cloning of T cell antigen receptor genes. 22 However, so far it has failed to achieve the status of an analytical tool: it has not been possible, by simple inspection of a difference library, to determine by how many species two mRNA populations differ. Use of hjac or hecc as vectors for difference library construction, together with 2D gels for their analysis, may redress this deficiency, as indicated in Fig. 5. Here, we were interested in mRNA species induced by lectin treatment of a T cell hybridoma. Comparison of 2D gels of the cell-free translation products of mRNA from treated and untreated cells did not reveal any obvious differences, so we could not use the direct, sib-selection approach previously successful in retrieving clones of interest. 2° Instead we constructed a difference library, enriching for induced cDNA species by subtractive hybridization with mRNA from untreated cells, as described above, prior to insertion into hecc. Figure 5B shows the translational readout of that difference library, while Fig. 5A shows the translation product of natural mRNA from untreated ceils. There are no spots that clearly occur on both gels: evidently subtractive hybridization efficiently removed most common species, allowing induced mRNAs, too sparse for their translation product to be detected after the translation of total mRNA, to dominate the difference library. This situation can be contrasted with the gels shown in Fig. 5C and D. These two gels show proteins expressed from two random sectors, both containing about 1000 clones, from an unenriched library of the mRNAs expressed in BW5147. In collaboration with Dr. I. Lefkovits we are currently attempting to catalog all of the genes detectably expressed in this cell line, by computer-assisted comparison of a large number of sectors like those shown in Fig. 5C and D. 21 mRNA species abundant in the cell will be represented in many of these sectors--as each random sector contains about 1000 clones, any mRNA species that makes up more than 0,1% of the total mRNA complement will have a good chance of being represented in any sector. Therefore, unlike the evaluation of the difference library in Fig. 5A and B, in this case one expects to see species common to both gels. Indeed, many such species can be discerned; 12 are indicated. NOTE ADDED IN PROOF. Since preparing this manuscript further observations, and a number of technical modifications that increase efficiency, have been made. They are listed here:
22 S. M. Hedrick, E. A. Nielsen, J. Kavaler, D. I. Cohen, and M. Davis, Nature (London) 308, 153 (1984).
IX 1] A
h PHAGE VECTORS
169
B
FIG. 5. Two-dimensional gels of cell-free translation products. (A) Polypeptides translated from mRNA of untreated BW5147 cells, using a nuclease-treated reticulocyte lysate, labeling with [35S]methionine. (B) Cell-free expression product of a hecc difference library, formed from concanavalin A-treated BW5147 cDNA, subtractively hybridized with mRNA from untreated cells. (C and D) Cell-free expression products of random sectors, each containing about 1000 clones, from a representative hecc library of BW5147 cDNA.
1. Strain TAP90, mentioned briefly above, is now used to grow recombinant phage in liquid culture for DNA preparation. DNA yields are severalfold increased over growth in LE392. However, it must be appreciated that there is no genetic selection for recombinants in this strain and therefore it should be used only for recombinants that have been through at least two rounds of selection in the P2 lysogen derivative of LE392. TAP90 can be obtained from the American Type Culture Collection (ATCC; Rockville, MD).
170
VECTORS FOR EXPRESSING CLONED GENES
[11]
2. Polyethylene glycol precipitates of phage from lysed liquid cultures are considerably cleaner if DNase and RNase are added immediately after lysis, before removal of bacterial debris. Add both enzymes to 1-2 tzg/ml of culture and shake at 37° for 10 min before addition of NaC1 to 1 M and centrifugation as described above. The precipitates formed by the addition to the supernatant of PEG to 10% can now, after pelleting, be dissolved directly in 10 mM EDTA, 25 mM Tris-HC1 (pH 8.6), 1% (w/v) SDS, containing 100/zg/ml proteinase K, then processed as described above. 3. Conditions now used are modified from those recommended by Gurevitch et al. 23for transcription of phage DNA. Twenty-microliter reactions contain 80 mM H E P E S / K O H , (pH 7.5), 10 mM DTT, 2 mM spermidine, 3 mM each of ATP, UTP, CTP and mVGpppG, 1 mM GTP, 0.5-1.0 /xg of phage DNA, 30 units T7 RNA polymerase, and 80 units RNase inhibitor. Incubation is at 38° for 3 hr. 4. Recent batches of Amersham reticulocyte lysate N90 have, alarmingly, proved almost entirely incapable of translating capped, ersatz mRNA generated by in vitro transcription of hecc recombinants, although these lysates were highly active on genuine mRNA. Fortunately the nuclease-treated reticulocyte lysate retailed by Ambion (Austin, TX) was capable of translating the same ersatz mRNA preparations, and the translation products yielded appeared normal on gel electrophoresis. The biochemical basis of this disturbing problem has not been determined; it is obviously imperative for users of the technology described here to screen several sources of lysate for activity on their in vitro transcripts, and to consider buying or making sufficient amounts of active lysate before embarking on any sizable series of experiments. Acknowledgments I thank I. Lefkovits and L. Kuhn for performing 2D gel analyses. St. Jude Children's Research Hospital is supported by Grant CA 21765 from the National Cancer Institute and by the American Lebanese Syrian Association (ALSAC).
23 V. V, Gurevich, I, D, Pokrovskaya, T. A. Obukhova, and S. A. Zozulya, Anal. Biochem. 195, 207 (1991).
[12]
EFFICIENT MUTAGENESIS IN PLASMIDS
173
[12] Site-Specific M u t a g e n e s i s in Plasmids: A G a p p e d Circle M e t h o d
By
B E R N D H O F E R a n d BIRGIT K ~ H L E I N
Site-specific mutagenesis is one of the key techniques in molecular genetics and protein engineering. The DNA segments to be mutagenized are usually contained in plasmids. The possibility of carrying out sitespecific mutagenesis with these plasmids directly, that is, without any subcloning steps, represents a substantial simplification and acceleration of this technique. A number of such procedures have in fact been described. However, the earlier methods either gave low yields of the desired mutation 1-7 or were not generally applicable as they needed a unique restriction site close to the target region of mutagenesis. 8-t° More recently, other laboratories devised techniques that circumvent these problems.l~-~4 Procedures using the polymerase chain reaction have also been developed tS-~s that are fast
R. B. Wallace, P. F. Johnson, S. Tanaka, M. Sch61d, K. Itakura, and J. Abelson, Science 209, 1396 (1980). 2 G. Dalbadie-McFarland, L. W. Cohen, A. D. Riggs, C. Morin, K. Itakura, and J. H. Richards, Proc. Natl. Acad. Sci. U.S.A. 79, 6409 (1982). 3 E. D. Lewis, S. Chen, A. Kumar, G. Blanck, R. E. Pollack, and J. L. Manley, Proc. Natl. Acad. Sci. U.S.A. 80, 7065 (1983). 4 B. A. Oostra, R. Harvey, B. K. Ely, A. F. Markham, and A. E. Smith, Nature (London) 304, 456 (1983). 5 S. M. Hollenberg, J. S. Lai, J. L. Weickmann, and T. Date, Anal. Biochem. 143, 341 (1984). 6 y. Morinaga, T. Franceschini, S. Inouye, and M, Inouye, Bio/Technology 2, 636 (1984). 7 K. Foss and W. H. McClain, Gene 59, 285 (1987). s W. Mandecki, Proc. Natl. Acad. Sci. U.S.A. 83, 7177 (1986). 9 G.-J, J. Chang, B. J. B. Johnson, and D. W. Trent, DNA 7, 211 (1988). 10 A. V. Bellini, F. de Ferra, and G. Grandi, Gene 69, 325 (1988), H B. Hofer and B. Kiihlein, Gene 84, 153 (1989). 12 M. Sugimoto, N. Esaki, H. Tanaka, and K. Soda, Anal. Biochem. 179, 309 (1989). 13 S. N. Slilaty, M. Fung, S.-H. Shen, and S. Lebel, Anal. Biochem. 185, 194 (1990). 14 D. B. Olsen and F. Eckstein, Proc. Natl. Acad. Sci. U.S.A. 87, 1451 (1990). J5 R. Higuchi, B. Krummel, and R. K. Saiki, Nucleic Acids Res, 16, 7351 (1988). z6 R. M. Nelson and G. L. Long, Anal. Biochem. 180, 147 (1989). r7 A. Hemsley, N. Arnheim, M. D. Toney, G. Cortopassi, and D. J. Galas, Nucleic Acids Res. 17, 6545 (1989). t8 M. Tomic, I. Sunjevaric, E. S. Savtchenko, and M. Blumenberg, Nucleic Acids Res. 18, 1656 (1990).
METHODS IN ENZYMOLOGY,VOL. 217
Copyright © 1993by AcademicPress, Inc. All rightsof reproductionin any form reserved.
174
MUTAGENESIS AND GENE DISRUPTION
[12]
but require an additional cloning step. This chapter details a method that yields average frequencies of mutants of around 60%. Principle of Method The fundamental prerequisite for site-specific mutagenesis is the presence of the DNA target region in single-stranded form to allow annealing of the mutagenic oligonucleotide. One way to achieve this (which has a number of advantages; see below) is the combination of two plasmid single strands of different origin to form a gapped plasmid. 4 This is shown in Fig. l, steps 2-4. An aliquot of the plasmid is cleaved with one or two restriction enzymes in such a way that the target segment for mutagenesis is cut out, and the nontarget segment is isolated. A second aliquot of the plasmid is linearized to allow physical separation of the two DNA strands in the following step. The respective cut must lie outside the target region to regenerate circular molecules in the subsequent reaction. Both double-stranded DNA (dsDNA) species are then mixed, "melted," and reannealed. As shown in Fig. l, this strategy in fact yields two complementary gapped circular molecules in equal amounts. As the mutagenic oligonucleotide will anneal to only one of them, the other will lead to wild-type progeny. To eliminate this "undesired" gapped circle, it is linearized and separated. Linearization is achieved by restriction endonuclease cleavage after selective annealing of an appropriate "restriction oligonucleotide" to its single-stranded region (Fig. 1, steps 5a and b). Subsequently, the remaining circular molecule can easily be separated from the linear species (which would be partially converted into circular molecules during the following elongation/ligation reaction) by agarose gel electrophoresis (AGE). To suppress the progeny from the wild-type strand of the heteroduplex, different mechanisms of biological selection have been exploited. 19-24The only one of these that is independent from any plasmid-based feature has been described by Kunkel et al. 21 for M13 vectors. The strand to be selected against is isolated from a d u t - u n g - host and consequently contains some dU residues. Such a DNA strand is preferentially degraded 19 S. Hirose, T. Kazuyuki, H. Hori, T. Hirose, S. Inayama, and Y. Suzuki, Proc. Natl. Acad. Sci. U.S.A. 81, 1394 (1984). 20 W. Kramer and H.-J. Fritz, this series, Vol. 154, p. 350. 21 T, A. Kunkel, J. D. Roberts, and R. A. Zakour, this series, Vol. 154, p. 367. 22 p. Carter, H. Bedouelle, and G. Winter, Nucleic Acids Res. 13, 4431 (1985). 23 M. A. Vandeyar, M. P. Weiner, C. J. Hutton, and C. A. Batt, Gene 65, 129 (1988), 24 j. Messing, this series, Vol. 101, p. 20.
[121 STEP 1
3/2
EFFICIENT MUTAGENESIS IN PLASMIDS
175
AQBtransform > reisolate A ~ B dut-g..ng-strain piasmid >
C cleave at A, f3 [ isolate non~ target fragment
C
|
I cleave at C A
4
5a
5b
6
denature and reanneal
B
J
anneal restriction a igonuc eotide D
l cleave at D
isolate gapped circ e
FIG. 1. Scheme illustrating the preparation of the plasmid substrate for mutagenesis. Normal and dU-containing DNA strands are distinguished by thick and thin lines, respectively. A, B, C, and D are restriction sites. The target segment for mutagenesis is indicated by zigzag lines. The linear products formed during reannealing are omitted for clarity. [Reprinted from B. Hofer and B. Kiahlein, Gene 84, 153 (1989) with permission.]
176
MUTAGENESIS AND GENE DISRUPTION
[12]
in strains that are wild type in this trait. The strategy (annealing of strands of different origin) used to form the gapped circle easily allows the combination of a dU-containing template strand with a normal complementary one (Fig. 1, steps 1-4). In addition, the use of a gapped circle, as compared to a completely single-stranded circle, has a number of advantages for mutagenesis. First, in vitro DNA synthesis is facilitated. Second, the major part of the DNA molecule is protected against unintentional annealing of the mutagenic oligonucleotide, which may lead to unwanted mutations (see Unintentional Mutations in General, below). Third, the probability of masking the target region by formation of stable intramolecular secondary structures is considerably reduced. This suggests, therefore, that the smaller one makes the gap, the better. One extreme is that the whole gap can be filled by the oligonucleotide. In this case unintentional annealing virtually cannot occur, and moreover the DNA polymerase reaction (which presumably reduces the mutant yield) is no longer necessary. However, if mutations are to be introduced into the same DNA segment at different sites, it is economical to choose a gap size that allows the same gapped circle to be used as a substrate for all mutagenesis experiments. Once the gapped circle has been isolated, the actual mutageneses are carried out in a conventional way. The phosphorylated oligonucleotide is annealed to the gapped circle, the gap is filled by a DNA polymerase, the remaining nicks are closed by a DNA ligase, and the resulting heteroduplex molecule is introduced into a bacterial cell.
Materials and Reagents
Strains CJ23625 [dutl ungl thil relA1/pCJl05 (Cmr)] BMH71-18mutS 26 [A(lac-proAB) thi supE mutS215::Tn10/F' l a d q ZAM15 proA +B +] DH127 [endA1 hsdRl7 (rk-, mk ÷) supE44 thil recA1 gyrA96 relA1] Media LB (per liter): 10 g Bacto-tryptone, 5 g Bacto-yeast extract, 10 g NaCI, pH adjusted to 7.5 with NaOH. For plates, 15 g Bacto-agar is added ,.5 j. Geisselsoder, F. Witney, and P. Yuckenberg, BioTechniques 5, 786 (1987). :6 B. Kramer, W. Kramer, and H.-J. Fritz, Cell 38, 879 (1984). 27 D. Hanahan, J. Mol. Biol. 166, 557 (1983).
[12]
E F F I C I E N T M U T A G E N E S I S IN P L A S M I D S
177
TB (per liter): Dissolve 12 g Bacto-tryptone and 24 g Bacto-yeast, extract in 900 ml water, add 4 ml glycerol, and autoclave. Dissolve 12.5 g K2HPO 4 and 2.31 g KH2PO 4 in 100 ml water and autoclave. Mix both solutions
Buffers and Solutions TE: 10 mM Tris-HC1 (pH 8.0), 1 mM ethylenediaminetetraacetic acid (EDTA) (pH 8.0) 8x HB (hybridization buffer): 1.5 M KC1, 0.1 M Tris-HCl (pH 7.5) 5 x KB (kinase buffer): 300 mM Tris-HCl (pH 8.0), 50 mM MgC12 , 0.5 mg/ml bovine serum albumin (BSA) 2 x MB (mutagenesis buffer): 23 mM Tris-HCl (pH 7.5), 9 mM MgC12, 4 mM dithiothreitol (DTT), 1.5 mM ATP, 0.8 mM each dNTP 5 x TMN: 200 mM Tris-HCl (pH 8.0), 50 mM MgCI2, 250 mM NaCI TENA: 40 mM Tris base, 1 mM EDTA, 20 mM sodium acetate, adjusted to pH 8.3 with acetic acid GMM (gel loading mix): 40% (v/v) glycerol, 50 mM EDTA (pH 8.0), 0.05% (w/v) bromphenol blue, 0.05% (w/v) xylene cyanol Solution I (minipreparation): 50 mM glucose, 25 mM Tris-HC1 (pH 8.0), 10 mM EDTA (pH 8.0) Solution II (minipreparation): 0.2 N NaOH, 1% (w/v) sodium dodecyl sulfate (SDS) Phenol solution: 227 g phenol, 100 ml I M Tris-HCl (pH 8.0), 12.5 ml water, 12.5 ml m-cresol, 0.5 ml 2-mercaptoethanol, 0.25g 8-hydroxyquinoline PCI solution: Phenol solution chloroform/isoamyl alcohol (50/48/2, v/v/v) CI solution: Chloroform/isoamyl alcohol (2/1, v/v) tRNA solution: 1/xg//xl tRNA (Bethesda Research Laboratories, Gaithersburg, MD) in TE
Enzymes T4 and T7 DNA polymerases (Pharmacia, Uppsala, Sweden), T4 DNA ligase (Boehringer Mannheim, Mannheim, Germany), T4 polynucleotide kinase (New England BioLabs, Beverly, MA), and restriction enzymes (Pharmacia, New England BioLabs, Bethesda Research Laboratories, and Boehringer Mannheim) are commercial preparations
Oligonucleotides Oligonucleotides are synthesized from phosphoramidite monomers on a "Gene Assembler" (Pharmacia) according to the instructions of the manufacturer and are purified by polyacrylamide gel electrophoresis.
178
MUTAGENESIS AND GENE DISRUPTION
[12]
Other Materials "Qiagen" columns (Qiagen, Chatsworth, CA) Method
Preparation of Gapped Circle Preparation ofdU-Containing Linear Plasmid. The plasmid to be mutagenized is isolated from a dut- ung- strain such as CJ236. It is cleaved at a preferably unique restriction site located outside the target region for mutagenesis. The DNA is purified and taken up in a buffer of low ionic strength or in water. This is necessary for formation of the gapped circles. For the plasmid isolation we initially used a standard CsC1 gradient protocol (see, for example, Ref. 28). When we tried a small-scale procedure, we encountered problems in formation of the gapped circles. However, when the minipreparation procedure was combined with a purification step using a Qiagen column, these problems could be overcome. Procedure 1. Prepare competent cells of a dut- ung- strain (e.g., CJ236) and transform them with the plasmid. We use the Hanahan procedure z7 with some slight modifications29: RbCI, DTT, and dimethyl sulfoxide are replaced by KC1, 2-mercaptoethanol, and dimethylformamide, respectively. 2. Screen a few transformants for the presence of intact plasmid: Isolate the plasmids by a quick minipreparation procedure (see, for example, Ref. 30), cleave them with a frequently cutting restriction enzyme and with the enzymes to be used for the mutagenesis, and check by AGE if the fragment patterns are as expected. 3. For a medium-scale preparation grow the transformed cells in a rich medium (e.g., TB) containing uridine (0.25 /zg/ml) and the appropriate antibiotic. The plasmid may be isolated using a standard CsC1 gradient protocol (see, for example, Ref. 28) or a Qiagen column as described by the supplier. 4. Linearize about 10 pmol of plasmid with an appropriate restriction enzyme. 5. Add EDTA to complex all divalent cations and purify the DNA by repeated extraction with the PCI solution. Remove the organic phase from the bottom of the tube to minimize the loss of plasmid. 28 D. Ish-Horowicz and J. F. Burke, Nucleic Acids Res. 9, 2989 (1981). 29 B. Hofer, Eur. J. Biochem. 167, 307 (1987). 30 D. S. Holmes and M. Quigley, Anal. Biochem. 114, 193 (1981).
[12]
EFFICIENT MUTAGENESIS IN PLASMIDS
179
6. After the last removal transfer the remaining aqueous phase to a new tube. Ethanol precipitate the plasmid and resolve it in 0.1 x TE. Preparation of Unmodified Nontarget Segment. Unmodified plasmid is isolated from a normal (i.e., dut + ung +) strain. The target segment is cut out with appropriate restriction enzyme(s). The nontarget segment is isolated by AGE and taken up in a buffer of low ionic strength or in water. For the isolation we use low melting gels and a modified version of the protocol of Wieslander. 31
Procedure 1. Isolate the unmodified plasmid as described in Preparation of dUContaining Linear Plasmid (above). 2, Digest about 10 pmol of the plasmid with the restriction enzyme(s) appropriate to cut out the target segment. 3. Add EDTA to complex all divalent cations and extract once with PCI solution. 4. To the aqueous phase add 0.25 vol of GMM and load on about 150 mm 2 of an appropriate low-melting-agarose gel. We generally use a horizontal submarine 0.8% (w/v) gel [14 (length) x 11 x 0.4 cm] prepared with TENA buffer containing 0.5 t~g/ml ethidium bromide. Run the gel in the same buffer at 70-100 V. 5. Mark the position of the nontarget fragment during brief illumination at 360 nm and cut out the gel piece. 6. Melt the gel piece at 65 ° . 7. Add 1 vol of phenol solution prewarmed to 65 °, vortex for 45 sec, and spin in an Eppendorf centrifuge at full speed for 2 min. Remove the organic phase, briefly recentrifuge the aqueous phase, and transfer it to a new tube. 8. Extract twice with 1 vol of PCI solution. 9. Extract once with 1 vol of CI solution, 10. Concentrate approximately twofold by extraction with 2.5 vol of l-butanol: vortex and spin for 30 sec each. I 1. Add 0.1 vol of 3 M sodium acetate, pH 4.8 (with acetic acid) and precipitate with 3 vol of ethanol, wash with 70% (v/v) ethanol, dry, and dissolve in 0. ! x TE. Formation of Gapped Circles. The dU-containing linear plasmid and the unmodified nontarget segment of the plasmid, both in water or in buffer of low ionic strength, are mixed. The duplexes are converted into single strands by heat, and these are reannealed by increasing the ionic strength followed by cooling to room temperature. This yields two complementary 31 L. Wieslander, Anal. Biochem. 98, 305 (1979).
180
MUTAGENESIS AND GENE DISRUPTION
1
gc~ lin~ fr~
2
[12]
3
~OC
-,~l lin
FIG. 2. Verification of gapped circle formation by AGE. Lane 1, mixture of the linearized plasmid (lin) and the nontarget fragment fir); lane 2, same as lane 1, but after the denaturation/ renaturation step (gc, gapped circle); lane 3, reference containing linear din) and open circular (oc) plasmid. [Reprinted from B. Hofer and B. Kiihlein, Gene 84, 153 (1989) with permission.]
gapped circle species (inseparable by AGE) as well as the linear starting products. In our experience, this step is susceptible to the purity and ionic strength of the DNA solutions, to the DNA concentration, and to minor changes in the melting/reannealing procedure. Therefore it is highly advisable to check the reaction by AGE (Fig. 2). The gapped circle migrates somewhat faster than the open circle (a by-product in most plasmid preparations, and which usually is a good marker), depending on the size of its gap. It may happen that the gapped circle band coincides with one of the linear DNA bands and may thus be hidden. This problem can always be resolved by altering the agarose concentration, which (along with other factors such as temperature) influences the relative mobilities of linear and circular DNA. We normally use a three- to fivefold excess of linear plasmid over nontarget segment, A poor formation of gapped circles was observed below as well as above certain DNA concentrations. In our hands, amounts of 1-5 pmol (2-10/zg) of linear plasmid in a final volume of 200 /~1 work well. The mixture of the gapped circles was found to be somewhat unstable even when stored at - 2 0 °. Presumably the gapped molecules are reconverted into the thermodynamically more stable starting products. Therefore, this mixture should be used within 1-2 weeks.
[12]
E F F I C I E N T M U T A G E N E S I S IN P L A S M I D S
181
The following procedure is a modification of the protocol of Kramer and Fritz. z°
Procedure 1. Mix 2.5 pmol of linear dU-containing plasmid, 0.75 pmol of unmodified nontarget segment, and H20 to yield a volume of 175/M. 2. Hold the mixture at 70 ° for 5 rain. 3. Add 25/zl of 8 × HB, prewarmed to 70 °, and continue the incubation at 70 ° for another 3 rain. 4. Cool the tube to room temperature in air (5 rain). Cleavage and Separation of Undesired Gapped Circle. To selectively cleave the undesired gapped circle, a "restriction oligonucleotide" is used that specifically anneals to it. This oligonucleotide contains one strand of a restriction site that must not occur outside the gap, but that may occur in the gap more than once. A 20-mer usually is of sufficient length to obtain complete cleavage at 370.9 Incubation with the restriction enzyme should not be extensively long as this may lower the yield of the desired gapped circle. Presumably some annealing of this DNA to the complementary gapped circular (or, after cleavage, gapped linear) species and/or to the oligonucleotide (via its restriction "half-site," which in most cases will be self-complementary) accounts for this decrease. Therefore, the enzyme concentration and incubation time given in the procedure should be regarded only as a guideline and may have to be determined empirically. Cleavage is analyzed by AGE. The intensity of the band representing the gapped circle(s) should have decreased by 50%. The remaining gapped circle is isolated from a preparative gel.
Procedure I. Mix 200/zl of the gapped circle solution with 50 pmol of the "restriction oligonucleotide." Supplement and adjust appropriately for the subsequent restriction digest. 2. Incubate for 5 min at 65 ° and for 5 min at room temperature (air). 3. Add 60 units of restriction enzyme and incubate at the appropriate temperature for 45 min. 4. Withdraw an aliquot for AGE; store the remainder at - 2 0 °. 5. If the reaction is complete, complex Mg 2+ with EDTA, extract once with PCI solution, ethanol precipitate (if the volume is too large, butanol concentrate or add carrier), and take up in 25/zl TE. 6. Add 0.25 vol of GMM and load on about 30 mm 2 of an appropriate low-melting-agarose gel. Run at 70-85 V. 7. Isolate the gapped circle as described in Preparation of Unmodi-
182
MUTAGENESIS AND GENE DISRUPTION
[12]
fled Nontarget Segment (above). Prior to elution add 4/xl tRNA solution to the gel piece as a carrier. (The tRNA remains in the mixture. There are no indications that it interferes with the mutagenesis procedure.) 8. Finally, take up the gapped circle in about 100/zl TE (corresponding to a DNA concentration of about 1 fmol//~l).
Oligonucleotide-Directed Mutagenesis Phosphorylation of Mutagenic Oligonucleotides: Procedure 1. Incubate 20 pmol of mutagenic primer in 15/zl KB, 10 mM DTT, 0.1 mM ATP with 1 unit T4 polynucleotide kinase for 30 min at 37 °. 2. Incubate for 10 min at 70 °, and briefly quench in ice. Use without purification.
Conversion of Gapped Circle into Covalently Closed Heteroduplex Circle. The mutagenic oligonucleotide is annealed to the isolated gapped circle, the remainder of the gap is filled by a DNA polymerase, and the nicks are closed by a DNA ligase. T4 DNA polymerase (at 37 °) consistently led to higher mutant yields than Klenow enzyme (at room temperature), as has also been reported by others. 25'32 The reason for this increase may be that the T4 enzyme has a reduced capacity to displace double-stranded structures. 33 However, T4 polymerase is more sensitive to secondary structures in the template. If this is a problem, addition of the T4 ssDNA-binding protein or of a small amount of Klenow enzyme is recommended. 34
Procedure 1. Mix 3.0/~1 of phosphorylated primer with 2.5/zl of isolated gapped circle, 1.0/zl H20, and 0.5/xl 8 × HB. 2. Incubate for 5 min at 65° and for 3 min at room temperature (air). Add 10.0/zl 2 × MB, 1.0/zl H20, 1/xl T4 DNA ligase (2 units//xl), and 1 /xl T4 DNA polymerase (1 unit//zl) 4. Incubate for I0 min at room temperature followed by 110 min at 37 °. 5. Add 2/xl 150 mM EDTA (pH 8.0). Transformations. Two subsequent transformations are carried out. In the first transformation, the strain must be wild type for dut and ung. It should also be deficient in mismatch repair, which seems to act efficiently .
32 W. Kramer, A, Ohmayer, and H.-J. Fritz, Nucleic Acids Res. 16, 7207 (1988). 33 N. G. Nossal, J. Biol. Chem. 249, 5668 (1974). 34 K. C. Deen, T. A. Landers, and M. Berninger, Anal. Biochem. 135, 456 (1983).
[12]
EFFICIENT MUTAGENESIS IN PLASMIDS
183
not only on incorrectly paired or unpaired bases, but also on bulge loops .35 Different strains 2°'35 are available that are deficient in the dam-instructed repair system; this seems to be responsible for the majority of repair events .35 As the primary transformants were found occasionally to contain progeny from both the normal (mutant) and the dU-substituted (wild-type) strand, a second transformation is carried out. The cells from the first transformation are therefore not spread on plates, but used to inoculate a liquid broth, from which plasmids are then isolated. A strain of choice is subsequently transformed with this plasmid mixture, and the cells are plated out. Procedure
1. Use 1-4/xl of the mixture of in vitro mutagenesis to transform 80 /xl of competent dut + ung + cells as described in Preparation of dUContaining Linear Plasmid (above). 2. After the 1-hr incubation in SOC medium, 27 use 200 ~zl of the cell suspension to inoculate 2 ml LB medium containing the appropriate antibiotic. Grow this culture overnight. 3. Isolate plasmid DNA from 1 ml of this culture by a fast minipreparation procedure (see, for example, Ref. 30) and take up in 50/xl 0.1 x TE. 4. Use 1/xl of a 1 : 100 dilution of this preparation to transform 20 pJ of competent cells (e.g., strain DH1). 5. Spread 3 /zl of the resulting 100 /zl on LB plates containing the appropriate antibiotic. A few hundred colonies are usually obtained. Screening for and Analysis of Mutants
Normally the analysis of three colonies of an individual mutagenesis experiment is sufficient to find at least one mutant clone. Therefore, DNA sequencing can be used for screening. However, if many mutageneses are carried out simultaneously, single-base sequencing or restriction analysis (when feasible) may be considered. Whatever screening method is used, we emphasize that it is insufficient to demonstrate only the presence of the intended mutation, and that it is absolutely necessary to sequence the whole region relevant for subsequent investigations to verify the absence of any unintentional mutation (see Unintentional Mutations, below). Isolation o f Plasmids. When RNase is omitted, plasmids prepared by the following minipreparation procedure 36 yield excellent sequencing 35 R. A. Fishel, E. C. Siegel, and R. Kolodner, J. Mol. Biol. 188, 147 (1986). 36 G. Morelle, Focus U , 7 (1989).
184
MUTAGENESIS AND GENE DISRUPTION
[12]
results. About 3 ml of each culture grown in LB medium was sufficient with our plasmids. A lower copy number may be compensated for by using a richer medium (e.g., TB).
Procedure 1. Pellet cells from 2.8 ml of each culture for 1 min at 8,000 rpm in an Eppendorf centrifuge; remove supernatants completely. 2. Resuspend pellets in 150/zl of ice-cold solution I by vortexing; leave at room temperature for 5 min. 3. Add 300 ~1 of solution II (freshly prepared), mix by inverting the tubes several times, and place on ice for 5 min. 4. Add 225/zl of 7.5 M ammonium acetate and mix by inverting the tubes several times; place on ice for 10 min. 5, Centrifuge at full speed for 5 rain, and carefully transfer about 600 /zl of the supernatants into new tubes. 6. To the supernatants add I vol of 2-propanol, mix, place on ice for 10 min, spin at full speed for I0 min, and discard supernatants. 7. To the pellets add ! ml of 70% (v/v) ethanol, vortex briefly, spin at full speed for 3 min, carefully remove the supernatants, and dry the pellets under vacuum for a few minutes. 8. Resolve the pellets in 100/zl 0.1 x TE. Double-Stranded DNA Sequencing. The plasmids isolated by the minipreparation procedure described above were denatured essentially as reported by Hattori and Sakaki. 37 If necessary, this step can be checked by AGE: the denatured plasmid migrates somewhat faster than the native form. The sequencing reaction was adopted from Tabor and Richardson 38 with a few modifications. When relatively high primer-to-template ratios were used (see Procedure, below), the signal-to-background ratio of the autoradiographs was usually indistinguishable from that obtained with CsCI gradient-purified DNA.
Procedure 1. Mix 80/xl of minipreparation and 20/zl of 1 N NaOH, 1 mM EDTA. Leave at room temperature for 5 min. 2. Mix with 40/zl 5 M ammonium acetate and 420/zl ethanol. Place in a - 7 0 ° freezer for 15 min, spin at full speed for 10 min (4°), and remove the supernatant. 3. Add 1 ml 70% (v/v) ethanol (-20°), vortex briefly, spin for 3 min 37 M, Hattori and Y. Sakaki, Anal. Biochem. 152~ 232 (1986).
3~ S. Tabor and C. C. Richardson, Proc. Natl. Acad. Sci. U.S.A. 84, 4767 (1987).
[12]
E F F I C I E N T M U T A G E N E S I S IN P L A S M I D S
185
as above, carefully remove supernatant, and dry the pellet under vacuum for a few minutes. 4. Resolve the pellet in 32 p~! H20. 5. Mix 7 t~l (about 0.25 pmol) of this solution with 2 t~l 5 × TMN and I/~1 primer (5 pmol). Incubate at 65° for 5 min, then at room temperature (air) for 3 rain. 6. Perform the primer elongations, for example, as described by Tabor and Richardson. 38 We take only half their dNTP concentration in the labeling step and use unmodified T7 DNA polymerase. Limitations and Modifications of the Method There are no limitations as far as the type (substitution, deletion, or insertion) and position of the mutation are concerned. However, three restriction sites (A, B, and C) are required to generate the gapped circles (Fig. 1). Sites A and B are likely to be found in any recombinant plasmid, because the cloning site(s) may serve this purpose. Site C (outside the target region), preferably unique but not necessarily, is also likely to be available in any plasmid. There is no absolute need for an additional site D to linearize the undesired gapped circle, as in most instances it should be possible to make use of either site A or site B for this purpose. In both gapped circles site A (or B) is incomplete. To our knowledge, there are no data in the literature on the substrate properties of incomplete (distinct from single-stranded) restriction sites. It is tempting to assume that such sequences are generally uncleavable; however, preliminary results indicate that this is not the case (see below). Three situations may be envisaged. 1. The site is not cleaved in both gapped circles. Therefore selective cleavage is possible by using a "restriction oligonucleotide" (or limited enzymatic elongation). 2. The site is cleaved in both gapped circles. In this case it cannot be used as site D. 3. The site is cleaved in only one of the two gapped circles. This gapped circle is then chosen to be the undesired one and can be selectively linearized without the need for a restriction oligonucleotide (or limited enzymatic elongation). The third case most likely applies only to sites produced by staggered cuts, as only then are the structures of the sites in the complementary gapped circles not identical. To illustrate this, when the gap has been generated using EcoRI and HindIII, the structures of the sites in the gapped circles (termed "gc + " and " g c - ") are as follows:
186
MUTAGENESIS AND GENE DISRUPTION
gc +
........ G AGCTT ........ ........ CTTAAG .......................... TTCGAA ........ / / /
gc-
[12]
/
........ GAATTC .......................... AAGCTT ........ ........ CTTAA A ........
The internucleotide linkages that would be cut by the endonucleases in the complete recognition sequences are indicated by slash marks (/). For each enzyme, the respective internucleotide linkages are located in the single-stranded segment of one gapped circle, but in the double-stranded segment of the other. Preliminary results indicate that HindlII is unable to cut either of its truncated sites, whereas EcoRI is able to cleave the site in g c - , but not in gc +. Therefore, when EcoRI is used, g c - is defined as the "undesired" gapped circle and can be linearized even without a restriction oligonucleotide. Unintentional Mutations
Mutant screening frequently was done by single-base or complete DNA sequencing. The application of this technique resulted in the accumulation of a substantial body of information on unintentional mutations (UMs).
Unintentional Mutations in General About one-half of the UMs observed can be explained by unintentional annealing of the mutagenic oligonucleotide in combination with the enzymatic activities of the in vitro system. An example is shown in Fig. 3. Evidently, even a single base pair at the 5' end of such a hybrid can be sufficient to stabilize the structure long enough to reach replication. The residual UMs comprised large deletions (hundreds of base pairs), which also might be mediated by template-primer interactions, and a few point mutations, which might be due to errors by a DNA polymerase (in vitro or in vivo). With some oligonucleotides the yield of UMs was high (about 40%) or even appeared to be coupled to the generation of the desired mutation. Thus it became a problem to obtain the intended mutation without the simultaneous introduction of the unintended one. In these cases an additional high-temperature incubation with the DNA polymerase prior to
[12]
187
EFFICIENT MUTAGENESIS IN PLASMIDS T
C C A @ A @ @ T T
T
@ A
@ T T C
T T
C TAAA©ACA@C
TA C@ @ATA A A C
A
I
3'
C @ CoC O
T,@ @C
AAT
TEMPLATE
@C@
AT
oT ~
. ojjJ'° @C TA TA TC@
@
ILPR33
@
T T @@C Co
C@ TA¢ T A Tc T @T @ T A A A T T C A A A A T OTA C @@Q;AAOT T C CcTOCTT
5' FIG. 3. Formation of an unwanted mutation by unintentional annealing of the primer ILPR33 to the template. The structure shown was calculated by a program of Zuker and Stiegler [M. Zuker and P. Stiegler, Nucleic Acids Res. 9, 133 (1981)] originally written for RNA, The G • T pair in the hybrid is indicated by a circle. Exonucleolytic degradation of the unannealed 3' end of the primer (8 nt), elongation, and finally ligation yielding a covalently closed circular (ccc) plasmid would explain the observed mutation.
elongation at room temperature (as originally described by Strauss e t al. 39 to increase the yield of site-specific deletions) proved to be most helpful. In these experiments we used the Klenow enzyme (under conditions de39 M. Strauss, C. H. Streuli, and B. E. Griffin, Gene 49, 331 (1986).
188
MUTAGENESIS AND GENE DISRUPTION
[12]
scribed by Kramer et al.4°). We have evidence, however, that the hightemperature step might also work with T4 polymerase (see below). After annealing of template and primer for 5 min at 68 ° (or, if lower, at the calculated Tm of the hybrid39), dNTPs, salts, and Klenow enzyme (0.75 units) were added and the incubation was continued at the same temperature for another 5 min. Only then was the mixture cooled to room temperature (in air for 5 min), after which 0.4 units of DNA polymerase and 5 units of DNA ligase were added, and the standard protocol was followed. This modified procedure was performed with three "problematic" primers. No UMs were found in 18 clones analyzed while the intended mutations were obtained at the same frequency as with the standard protocol. Experiments with 5'-labeled primers demonstrated that all unannealed molecules were partially degraded by the 3'-exonucleolytic activity of the Klenow enzyme during the high-temperature (68 °) incubation. This gives a rationale for the observed suppression of UMs formed by unintentional annealing. So far, we have not needed to use the additional high-temperature step when mutagenesis reactions were carried out with T4 DNA polymerase. If necessary, it should be possible to apply the T4 enzyme successfully in the high-temperature protocol. A pilot experiment as described in the preceding paragraph, but under T4 mutagenesis conditions, showed that after the 68 ° step the selective degradation of unannealed 28-mer primers was, as may be expected, even more pronounced than when the Klenow enzyme was used. Unintentional M u t a t i o n s in Insertions
When mutant DNAs were sequenced that originated from experiments with oligonucleotides designed to generate insertions, as many as 17 of a total number of 47 (36%) showed sequence deviations in the regions originating from the mutagenic primers. We observed substitutions and deletions of one or two adjacent or nonadjacent nucleotides (nt). Remarkably, all of these sequence errors were located in the insertions (varying from 27 to 45 nt in length), that is, in those parts of the primers that did not base pair with the template. They are probably due to errors in chemical synthesis and not to some biological loop "repair" mechanism. When we used one of these primers in a modified mutagenesis procedure (currently under investigation) in which the wild-type strand was degraded beyond 40 W. Kramer, V. Drutsa, H.-W. Jansen, B. Kramer, M. Pflugfelder, and H.-J. Fritz, Nucleic Acids Res. 12, 9441 (1984).
[13]
HIGH-EFFICIENCY
SITE-DIRECTED
MUTAGENESIS
189
the position of the loop prior to transformation, unintentional mutations again were found in some of the inserts. 41 Results reported by Clackson and Winter42 also argue against a biological phenomenon. They replaced 383-nt-long DNA segments by site-directed mutagenesis using polymerase chain reaction (PCR)-generated primers. Despite the formation of 383-ntlong (interior) loops on annealing of the polynucleotides, all 14 positive clones sequenced were found to be correct. Acknowledgments The authors wish to thank H. Bl6cker, R. Frank, and co-workers for oligonucleotide synthesis, P. Artelt, J. Hoppe, and H.-J. Fritz for the gifts of vectors and strains, and R. Brownlie for linguistic advice. This work was supported by the Bundesministerium fiir Forschung und Technologie through Grant No. 03 8706 9. 4J R. Wefel and B. Hofer, unpublished observations (1991). 4: T. Clackson and G. Winter, Nucleic Acids Res. 17, 10163 {1989),
[13] S i t e - D i r e c t e d M u t a g e n e s i s of S i n g l e - S t r a n d e d a n d Double-Stranded DNA by Phosphorothioate Approach
By DAVID B. OLSEN, JON R. SAVERS, and FRITZ ECKSTEIN Introduction
Oligonucleotide-directedmutagenesis allows the introduction of almost any precisely defined mutation into a cloned, sequenced gene. The mutation may comprise single or multiple mismatches or it may involve the insertion or deletion of a large number of bases. There are a number of methods described in the literature for the efficient production of site-directed mutations. The gapped duplex,l uracilcontaining template, 2 and coupled primer approaches3 have all been used to improve the basic method described in detail by Zoller and Smith. 4-6 However, these methods are limited to protocols using single-stranded vectors, involve the transfection of heteroduplex DNA (resulting in the W. Krammer and H.-J. Fritz, this series, Vol. 154, p. 350. z T. A. Kunkel, J. D. Roberts, and R. A. Zakour, this series, Vol. 154, p. 367. P. Carter, this series, Vol. 154, p. 382. 4 M. J. Zoller and M. Smith, Nucleic Acids Res. 10, 6487 (1982). -~M, J. Zoller and M. Smith, this series, Vol. 100, p. 468. M. J. Zoller and M. Smith, this series, VD1. 154, p. 329.
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
190
MUTAGENESIS AND GENE DISRUPTION
[13]
need for specialized E s c h e r i c h i a coli strains such as mismatch repairdeficient cells, d u t - u n g - , etc.), and very often require plaque purification steps. The original phosphorothioate-based mutagenesis method 7 has undergone a number of improvements, s-t° and has several advantages over the previously mentioned methods. The fundamental difference is that selection against the wild-type sequence is carried out in vitro. Therefore, special cell lines and plaque purification are avoided. Very high mutational efficiencies are obtainable, on the order of 70-90%, 9A° which allows for direct genotypic screening by DNA sequencing) 1 In addition, random mutagenesis procedures, such as those developed by Knowles and coworkers, ~2:3 are also possible due to the high efficiency that is essential for the productive application of such protocols. Finally, the phosphorothioate approach is not limited to the use of single-stranded or phagemid 14 DNA but has been extended to double-stranded DNA vectors. ~°'~5
Principle of Phosphorothioate-Based Mutagenesis Methodology for Single-Stranded DNA The phosphorothioate-based oligonucleotide-directed mutagenesis method exploits the observation that several restriction endonucleases cannot linearize DNA containing certain phosphorothioate internucleotidic linkages. 8'16-19The first step of the procedure involves annealing of a mismatch oligonucleotide primer to the (+)strand of a single-stranded
7 j. W. Taylor, J. Ott, and F. Eckstein, Nucleic Acids Res. 13, 8765 (1985). s K. L. Nakamaye and F. Eckstein, Nucleic Acids Res. 14, 9679 (1986). 9 j. R. Sayers, W. Schmidt, and F. Eckstein, Nucleic Acids Res. 16, 791 (1988). 10 D. B. Olsen and F. Eckstein, Proc. Natl. Acad. Sci. U.S.A. 87, 1451 (1990). u L. Serrano, A. Horovitz, B. Avaron, M. Bycroft, and A. R. Fersht, Biochemistry 29, 9343
(1990). 12 j. D. Hermes, S. M. Parekh, S. C. Blacklow, H. Koester, and J. R. Knowles, Gene 84, 143 (1989). 13 j. D. Hermes, S. C. Blacklow, and J. R. Knowles, Proc. Natl. Acad. Sci. U.S.A. 87, 696 (1990). 14 j. Vieira and J. Messing, this series, Vol. 153, p. 3. i5 D. B. Olsen and F. Eckstein, in "Directed Mutagenesis: A Practical Approach" (M. J. McPherson, ed.), p. 83. IRL Press, Oxford, England, 1991. 16 j. W. Taylor, W. Schmidt, R. Cosstick, A. Okruszek, and F. Eckstein, Nucleic Acids Res. 13, 8749 (1985). 17j. R. Sayers, D. B. Olsen, and F. Eckstein, Nucleic Acids Res. 17, 9495 (1989). 18 D. B. Olsen, G. Kotzorek, and F. Eckstein, Biochemistry 29, 9546 (1990). 19 D. B. Olsen, G. Kotzorek, J. R. Sayers, and F. Eckstein, J. Biol. Chem. 265, 14389 (1990).
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
A
B
C
D
19l
E
SCHEME[. Schematic representationof the phosphorothioate-based mutagenesis method. Single-stranded DNA, annealed with a mismatch primer [ t , position of the mismatch nucleotide(s)] (A), is converted to RFIV DNA using T7 DNA polymerase, T4 DNA ligase, and dNTPaS mix (B). The region containing phosphorothioate internucleotidic linkages in the newly synthesized ( - )strand is drawn with bold lines. The wild-type ( + )strand is specifically hydrolyzed by reaction with a restriction endonuclease resulting in a nicked DNA product (C). The nick is taken as the starting point for digestion by either a 3' ~ 5'- or a 5' ---, 3'-exonuclease (D). A fully complementary homoduplex RFIV molecule, which is ready for transformation, is generated on repolymerization (E).
circular phage DNA (Scheme IA). The primer is extended by T7 DNA polymerase using a mixture of three deoxynucleoside triphosphates and the Sp-diastereomer of a deoxynucleoside 5'-O-(l-thiotriphosphate) (dNTPo~S) such as dCTPaS (Fig. 1). The resulting ( - )strand of the newly synthesized RFIV DNA contains phosphorothioate groups (see bold lines in Scheme IB). This strand asymmetry is exploited by reaction with a restriction enzyme (e.g., NciI 8) that hydrolyzes only the wild-type (+)strand (Scheme IC). The resulting nick is converted to a gap by reaction with an exonuclease (Scheme ID). The gapped DNA is repolymerized using the mutant strand as the template, resulting in the formation of a mutant homoduplex with the mutant sequence present in both strands (Scheme IE). The DNA can be transformed into any E. coli host strain.
NH 2
-0-,,
o
o
o
II
II
!
P~O
I
O.
P ~ O ~ P ~ O
I
O-
....
t
S OH
FIG. 1. Structure of the Sp-diastereomer of deoxycytidine 5'-O-(1-thiotriphosphate), dCTPaS. The sulfur atom replaces a nonbridging oxygen atom of the a-phosphorus.
192
MUTAGENESIS AND GENE DISRUPTION
[13]
Experimental
Media 2YT medium: 16 g tryptone, l0 g yeast extract, 5 g NaC1 per liter; autoclave B-broth soft agar: 0.5 g tryptone, 0.4 g NaCI, 50/zl 1% (v/v) B~ solution, 0.3 g agar in 50 ml H20; autoclave Bx solution: l0 mg/ml thiamin hydrochloride
Reagents TE buffer: 10 mM Tris-HCl (pH 8), l mM ethylenediaminetetraacetic acid (EDTA); autoclave NTE buffer: 100 mM NaCI, I mM EDTA, 10 mM Tris-HC1 (pH 8); autoclave DNA buffer: 20 mM NaCI, 1 mM EDTA, 20 mM Tris-HCl (pH 8); autoclave 4 x dNTP mix: l0 mM dATP, l0 mM dCTP, 10 mM dGTP, 10 mM dTTP; sterile filter dCTPaS mix: 5 mM dATP, 5 mM dCTPaS, 5 mM dGTP, 5 mM dTTP; sterile filter dGTPaS mix: 5 mM dATP, 5 mM dCTP, 5 mM dGTPaS, 5 mM dTTP; sterile filter Buffer A (10 x ): 100 mM MgCI z , 50 mM dithiothreitol (DTT), 500 mM Tris-HC1 (pH 8) Buffer B (10 x): 1 M Tris-HCl (pH 8), 1 M NaC1; autoclave Buffer C (10 x ): 70 mM MgC12, 50 mM DTT, 100 mM Tris-HCl (pH 8), 600 mM NaC1; sterile filter. Prepare immediately prior to use Buffer D (10 x ): 80 mM MgClz, 400 mM NaCI, 500 mM Tris-HCl (pH 7.4); autoclave Isopropyl-/3-o-thiogalactopyranoside (IPTG) solution: 30 mg IPTG in 1 ml H20; sterile filter 5-Bromo-4-chloro-3-indolyl-3-galactoside) (X-Gal) solution: 20 mg X-Gal in 1 ml deionized dimethylformamide (do not sterilize!) EcoRI nicking buffer (10 x ): 4 mM CoClz, 1 M NaC1, 1 M Tris-HC1 (pH 7.4) Ethidium bromide solution: 0.5 mg/ml
Materials and Enzymes The Sp-diastereomers of the deoxynucleosides [5'-O-(1-thiotriphosphates)] were purchased from Amersham (Amersham, England) or syn-
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
193
thesized according to the procedure of Ludwig and Eckstein) ° Qiagen (Diisseldorf, Germany) tip 500 was used for the isolation of plasmid DNA. Nitrocellulose filters (SMl1336, 13 mm in diameter, 0.45-/xm pore size) were supplied by Sartorius. Filter units were from Millipore (Bedford, MA). Centricon-100 filtration units were obtained from Amicon Corporation (Danvers, MA). The enzymes T7 DNA polymerase, T4 polynucleotide kinase, and T7 exonuclease were obtained from United States Biochemicals (Cleveland, OH). All restriction endonucleases and exonuclease III were purchased from New England Biolabs (Beverly, MA) and E. coli DNA polymerase I was from Boehringer Mannheim (Mannheim, Germany). T4 DNA ligase 16and T5 exonuclease 21were prepared as previously described. Preparation of Single-Stranded Template DNA The first step toward successful completion of a mutagenesis experiment is effective growth and isolation of phage DNA. One of the biggest mistakes that can be made is to attempt site-directed mutagenesis with DNA that is not free from RNA or DNA fragments that are capable of acting as primers in the polymerization reaction. Even the novice can prepare sufficiently pure DNA without the need to perform RNase treatments or cesium chloride density gradient purifications. The method we recommend (procedure 1) is given below and includes two polyethylene glycol phage precipitation steps that are important factors in the preparation of suitable template DNA. Although the procedure is stretched over several days, it requires only a minimal amount of time during the first 3 days. When performing this procedure, it is important to avoid allowing phage particles to contact any of the solutions used for transformation of the mutated DNA.
Procedure 1: Preparation of Template DNA The phage carrying a sequenced insert must be present in a singlestranded DNA vector such as one of the M13 vectors characterized by Messing. 22
Day 1 1. Plate out the phage so as to give single plaques. 2. Prepare an overnight culture of, for example, SMH50 or TG1 cells 20 j. Ludwig and F. Eckstein, J. Org. Chem. $4, 631 (1989). 21 j. R. Sayers and F. Eckstein, J. Biol. Chem. 265, 18311 (1990). 22 j. Messing, this series, Vol. 101, p. 20.
194
MUTAGENESIS AND GENE DISRUPTION
[13]
in 3 ml 2YT medium by picking a colony from a glucose-minimal medium plate. (If starting with DNA, follow procedure 12 for the transformation of competent cells).
Day 2 1. Prepare fresh cells by adding I drop of overnight culture into 3 ml fresh 2YT medium and incubate at 37° for 3 hr in a shaker. 2. Prepare a phage solution by picking a single plaque into 100/xl of fresh cells and incubate overnight at 37 °. 3. Set up another 3-ml overnight cell culture.
Day 3 1. Inoculate 100 ml 2YT medium (in a 250-ml flask) with l ml of fresh cells (prepared as described above) and grow at 37°, with shaking, to an A660of 0.3. 2. Add the phage solution and continue incubation for 5 hr. 3. Transfer the solution to centrifuge tubes and pellet the cells by centrifugation for 20 min at 23,000 g in a Sorvall (Norwalk, CT) centrifuge using a GSA rotor. 4. Immediately decant the supernatant and add 1/5 vol of 20% (v/v) polyethylene glycol (PEG) 6000 in 2.5 M NaC1. Allow the phage to precipitate for 30 min (or overnight) at 4 °.
Day 4 1. Centrifuge at -3500 g in a Sorvall centrifuge using a GSA rotor for 20 min at 4 °. Discard the supernatant and remove traces of liquid with a tissue or drawn-out pipette. 2. Add 10 ml TE buffer and resuspend the phage pellet. 3. Centrifuge at -3500 g for 20 rain at 4 °. Transfer the phage containing supernatant to a clean centrifuge tube. 4. Add 2.2 ml 20% (v/v) PEG in 2.5 M NaCI. Precipitate at 4 ° for 30 min. Centrifuge at -3500 g for 20 min at 4 °. 5. Discard the supernatant, remove any excess liquid using a drawnout pipette, and then dissolve the phage pellet in 500/.d NTE buffer and transfer to a sterile microcentrifuge tube. 6. Add 200/zl of buffer-equilibrated phenol, vortex for 30 sec, and spin briefly in a microcentrifuge. Transfer the aqueous (upper) layer to a new microcentrifuge tube. 7. Repeat step 6. 8. Add 500/xl H20 saturated diethyl ether, vortex for 30 sec, and spin briefly in a microcentrifuge. Discard the upper (ether) layer. Repeat the process three more times.
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
195
9. Add 50 ~l 3 M sodium acetate (pH 6), vortex, and divide the solution equally between two microcentrifuge tubes. 10. Add 700/A absolute ethanol to each tube, and cool to - 7 0 ° in a dry ice/2-propanol bath for 60 min. Centrifuge for 5 min at 14,000 rpm in a microcentrifuge. 11. Discard the supernatant, add 700/A 70% (v/v) ethanol, and invert the tube to drain off the solvent, taking great care not to dislodge the pellet. 12. Label the tubes X and Y and add 50 IA DNA buffer to tube X. Resuspend the pellet by vortexing and transfer the buffer containing DNA from tube X to tube Y. 13. Add 50 ~1 of DNA buffer to tube X, vortex, centrifuge briefly, and transfer the contents to tube Y. Vortex to resuspend the pellet. 14. Take a 10-/.d sample, dilute to 1000/A, and determine the optical density on an ultraviolet (UV) spectrometer at 260 and 280 nm in a 1-mi quartz cuvette. The ratio of A260/280should be 1.8 or higher; if not, repeat the phenol extraction and the following steps. One A260unit corresponds to - 3 7 / ~ g single-stranded DNA. Keep this sample as a standard for gel analysis (14/~1 diluted with stop mix). We highly recommend that a self-priming test (procedure 2, below) be performed on all newly isolated single-stranded DNA. After the test, the DNA can be analyzed by agarose gel electrophoresis. If significant amounts of polymerized material are observed we recommend that procedure 1 be repeated. Figure 2 shows the results after each stage of a typical single-stranded DNA mutagenesis experiment. The novice should compare results after each stage of the mutagenesis procedure with the results presented in Fig. 2.
Procedure 2: Self-Priming Test for Single-Stranded DNA 1. To a 1.5-ml sterile microcentrifuge tube, add Buffer B, 2.5 p.l Single-stranded DNA template, 5/~g 2. Adjust the final reaction volume to 23 ~l with H20. 3. Briefly vortex and spin down the solution. 4. Incubate at 70 ° for 5 min in a hot water bath. Transfer immediately to a heating block at 37 ° and leave for 20 rain. Place on ice. Then add Buffer A, 3.5 ~1 4 × dNTP mix, 2 pA ATP (10 mM), 5 pA
196
MUTAGENESIS AND GENE DISRUPTION
[13]
/ FIG. 2. Analysis of single-stranded DNA mutagenesis intermediates by agarose gel elec-
trophoresis.
DNA polymerase I, 5 units T4 DNA ligase, 5 units 5. Bring the volume of the solution to 35/zl using sterile H20. 6. Briefly vortex and spin down the solution. 7. Incubate at 37° for 2 hr. Remove a 2-/zl sample for gel analysis. This analysis should be carried out with a control reaction containing an appropriate primer that can anneal to the single-stranded DNA.
Phosphorothioate-Based Mutagenesis Using Single-Stranded DNA Vectors
Mutant Oligonucleotide The sequence of the mutant oligonucleotide determines how the target DNA sequence is to be mutated. Different types of changes such as transition or tranversion mutations as well as insertions and deletions are all possible with the phosphorothioate-based procedure. As previously
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
197
pointed out, 23 this method provides a distinct advantage when performing insertion mutagenesis because the selection step against the wild-type strand occurs in vitro. Preferentially, the oligonucleotide should have 6-7 bases on the 3' end to protect it from the 3' --~ 5' exonuclease proofreading activity of the polymerase. 24 For single- or double-base mismatches we routinely use oligomers of 18-22 nucleotides in length with the mismatch(s) positioned toward the center. Another important concern is that the oligonucleotide does not contain a high degree of self-complementarity. This could cause problems in the annealing step due to self-association. Finally, the primer should not normally contain a recognition site for the restriction enzyme that is to be used in the nicking reaction. Such a site, if present, would lead to the linearization of the DNA during the nicking reaction as the primer does not contain any phosphorothioate groups. However, it is possible to use a primer with such a recognition site provided that the primer is chemically synthesized with phosphorothioate groups at the positions required to protect it from endonuclease-catalyzed hydrolysis. 25 Procedure 3 below provides a simple procedure for the phosphorylation of the mutant oligomer. A phosphorylated primer is required so that ligation can occur after complete synthesis of the mutant strand resulting in the conversion of RFII to RFIV DNA.
Procedure 3: Phosphorylation of Mismatch Oligonucleotide 1. Add the following to a sterile 1.5-ml microcentrifuge tube: Buffer A (10x), 3.5/xl ATP (I0 mM), 3/xl Oligonucleotide primer (stock of 5 A260units/ml for an oligomer of 18-24 bases), 2/xl 2. Bring the volume to 34/zl using sterile H20. 3. Add 5 units of polynucleotide kinase. 4. Briefly vortex and spin down the solution. 5. Incubate in a heating block at 37 ° for 15 min and then heat inactivate the enzyme at 70° for 10 min in a water bath. Store on ice.
23 T. A. Kunkel, in "Nucleic Acids and Molecular Biology" (F. Eckstein and D. M. J. Lilley, eds.), Vol. 2, p. 124. Springer-Verlag, Berlin, 1988. 24 S. Tabor, H. Huber, and C. C. Richardson, J. Biol. Chem. 252, 16212 (1987). 25 R. P. Iyer, L. R. Phillips, W. Egan, J. B. Regan, and S. L. Beaucage, J. Org. Chem. 55, 4693 (1990).
198
[13]
MUTAGENESIS AND GENE DISRUPTION TABLE I RESTRICTION ENDONUCLEASES UNABLE TO HYDROLYZE PHOSPHOROTHIOATECONTAINING DNA
Enzyme A v a I b'c AvalI b BamHI BanII b EcoRI EcoRV b FspI HindII
HindIII KasI NciI b,c PstI b
PvuI b PvulI SacI SmaI
DNA ~ M13mp2 qbX174 M13mp2 M 13mpl 8 Ml3mp2 M 13mp 18 M 13mp 18 M13mpl8 Ml3mpl8 M13mp2 M13mp9 M13mpl8 M 13mp 18 M 13mp2 M 13mp2 M13mpl8 M 13mp9 M13mpl8 pUC19 M13mp2 Ml3mpl8 M13mpl8 M13mpl8
Analog used for polymerization
Ref.
dCTPaS dTTP~S dGTPczS dATPc~S/dGTPaS dCTP~S/dNTPaS f
d d d e d, f, g
dCTPaS/dGTPaS f
dATPaS/dGTPczS dATPaS dGTP~xS dGTPczS
e e, g, h i d, i
dATPaS/dTTPo~S dGTPczS dCTPczS
e i d, j
dGTPaS
dCTPaS dCTPczS/dGTPt~S dCTPaS/dGTPaS
dGTPaS k
d, g, j
d e e l
a The initial nicking conditions were determined using the DNA vectors listed. We recommend that all nicking reactions be carried out according to the buffer and incubation conditions given in the original reference. o Most consistent results have been obtained using these enzymes. c This restriction endonuclease recognizes a degenerate recognition sequence and therefore incorporation of a different phosphorothioate might be required for the inhibition of hydrolysis with different DNA vectors. a j. W. Taylor, W. Schmidt, R. Cosstick, A. Okruszek, and F. Eckstein, Nucleic Acids Res. 13, 8749 (1985). e j. R. Sayers, D. B. Olsen, and F. Eckstein, Nucleic Acids Res. 17, 9495 (1989). f B a n l I recognizes the sequence 5'-GPuGCPy/C-3'. Our results [D. B. Olsen, G. Kotzorek, J. R. Sayers, and F. Eckstein, J. Biol. Chem. 265, 14389 (1990)] indicate that to inhibit this enzyme a phosphorothioate must be at the two positions designated by asterisks in the following sequence: 5'-GPuGCPy*C*N-3'. For M13mp2, rap7, mp8, mp9, or mpl0, the 3'-N is a cytosine in the ( - ) strand and therefore is protected by the presence of only dCMPS. However, Ml3mpl8 has a G in the ( - ) strand 3' to the recognition sequence and therefore the DNA must contain dCMPS and dGMPS groups.
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
199
Preparation of RFIV Heteroduplex DNA Annealing. For most mutations two molar equivalents of primer are annealed to the target sequence in a high-salt buffer, as described in Procedure 4. Procedure 4: Annealing of Primer to Template DNA 1. Add the following to a sterile microcentrifuge tube: Buffer B (10 ×), 3.5 ~1 Phosphorylated primer solution from procedure 3, 6 ~1 Single-stranded DNA template (from procedure 1, typically 2-5/xg//~l), 10/xg 2. Adjust the final reaction volume to 35/xl with H20. 3. Briefly vortex and spin down the solution. 4. Incubate at 70 ° for 5 min in a hot water bath, transfer immediately to a heating block at 37°, and incubate for 20 min before placing on ice. Polymerization. The choice of deoxynucleoside phosphorothioate for the polymerization reaction is dependent on which restriction enzyme is to be used in the subsequent nicking reaction (see Table I for a list of appropriate restriction endonucleases). The restriction enzymes NciI and AvaI have been used most extensively and both require the incorporation of dCMPS groups into M13mpl8 DNA to limit the hydrolysis to the wildtype strand. Native T7 DNA polymerase is the enzyme of choice for extension of the mutant oligonucleotide and complete synthesis of the mutant ( - )strand. This enzyme has several advantages over other DNA polymerases in that it is very processive, it does not strand displace, and pure enzyme is commercially available. In addition, this enzyme efficiently incorporates dNMPS analogs and the polymerization reaction is normally complete after short incubation periods. Alternatively, the Klenow fragment of DNA polymerase I yields acceptable results when used at 16° overnight.
g D. B. Olsen and F. Eckstein, Proc. Natl. Acad. Sci. U.S.A. 87, 1451 (1990). h D. B. Olsen, G. Kotzorek, and F. Eckstein, Biochemistry 29, 9546 (1990). i C. Krekel and J. R. Sayers, unpublished observations. J K. L. Nakamaye and F. Eckstein, Nucleic Acids Res. 14, 9679 (1986). k Nicking using Sinai requires 40/zg/ml ethidium bromide in the reaction in addition to dGMPS at the site of cleavage. J. R. Sayers, W. Schmidt, A. Wendler, and F. Eckstein, Nucleic Acids Res. 16, 803 (1988).
200
MUTAGENESIS AND GENE DISRUPTION
[13]
Procedure 5: Polymerization Reaction
1. Add the following reagents to the template/primer mixture (procedure 4) after the annealing mixture has been cooled on ice for 10 min: dCTPaS, 8/xl ATP (10 mM), 8/zl Buffer A (10×), 8/zl T7 DNA polymerase, 26 6 units T4 DNA ligase, 10 units 2. Adjust the volume to 80/zl with sterile H20. 3. Briefly vortex and spin down the solution. 4. Incubate for 1 hr at 37 °. Heat inactivate at 70° for 10 min and remove a 2-/zl sample for agarose gel analysis (see Fig. 2). Removal o f Wild-Type Single-Stranded DNA. The nitrocellulose filtration step described below is designed to remove any unpolymerized singlestranded DNA that remains after the polymerization reaction. Even though our procedure uses an excess of oligonucleotide for priming, some single-stranded DNA usually remains after the reaction, which can greatly reduce mutational efficiency. Procedure 6: Nitrocellulose Filtration z7
1. Using forceps place the rubber seal and two nitrocellulose filters (do not use autoclaved filters) in the female end of the filter housing. 2. Carefully apply 40/zl 500 mM NaCI to moisten the filter disks and then assemble the unit. 3. Attach a 2-ml disposable syringe to the outlet side of the filter unit using a l-cm length of silicone tubing. 4. Add 6/xl 5 M NaCI to the polymerization reaction. Mix and apply to the inlet side. 5. Slowly draw the sample through the filter unit using the syringe plunger. If necessary tap the housing gently to collect the filtrate. 6. Add 50/zl 500 mM NaC1 to the top of the filter unit and draw the wash through. 7. Carefully remove the filter unit and transfer the filtrate into a fresh, sterile microcentrifuge tube. 8. Rinse the syringe with 50/zl 500 mM NaC1 and combine with the filtrate. 26 It is important to add the T7 DNA polymerase after the nucleotide mix; otherwise, the strong proofreading activity associated with the enzyme may digest the mutant oligonucleotide. 27 The phosphorothioate-based mutagenesis kit supplied by the Amersham Corporation contains filter units that are operated by centrifugal force instead of a syringe system.
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
201
9. Add 400/zl cold absolute ethanol, mix, and place at - 7 8 ° for 15 min. 10. Spin in a microcentrifuge for 15 min at 14,000 rpm at room temperature. 11. Discard the supernatant (a small pellet of salt/DNA should be visible). 12. Carefully add 400/zl of 70% ethanol and invert the tube. Check that the pellet has not been dislodged. Open the cap to release the liquid, being careful not to disturb the pellet. 13. Remove the remaining amount of liquid using a Speed-Vac concentrator (Savant Instruments, Inc., Farmingdale, NY) for 2-3 min at room temperature.
Preparation of Mutant Homoduplex Strand-Selective Hydrolysis of Wild-Type DNA. There are a number of restriction endonucleases that can be used to hydrolyze the unmodified (nonphosphorothioate containing) strand of the mutant heteroduplex DNA (Table I). Below we have provided conditions for nicking reactions with the enzymes NciI and AvaI. After the nicking reaction is complete, we recommend that the user analyze the reaction products by agarose gel electrophoresis. There should be no RFIV DNA visible on the gel. Procedure 7: Restriction Endonuclease Nicking of Heteroduplex DNA NciI reaction 1. Resuspend the pellet (from procedure 6) in 190/zl H20 and add Buffer A (10 x ), 25/zl NciI, 120 units 2. Bring the volume to 250 txl with sterile H20. 3. Briefly vortex and spin down the solution. 4. Incubate at 37° for 90 min. Heat inactivate the enzyme at 70 ° for 10 min. Keep a 6-/zl sample for gel analysis (see Fig. 2). Alternatively, the DNA may be nicked with the enzyme AvaI. 1. Resuspend the DNA pellet in 160/xl H20 and add Buffer C (10 x ), 25/zl AvaI, 70 units 2. 3. 4. a 6-/zl
Bring the volume to 250/~1 with sterile H20. Briefly vortex and spin down the solution. Incubate at 37 ° for 180 rain. Heat inactivate at 70° for 10 min. Take sample for gel analysis (see Fig. 2).
202
MUTAGENESIS AND GENE DISRUPTION
[13]
Hydrolytic digestion of wild-type DNA by exonuclease digestion. The 5'--* 3' activity of T7 exonuclease gaps nicked double-stranded DNA and removes almost all the nicked wild-type strand under normal gapping conditions. Partially gapped DNA species are not detected by agarose gel analysis, indicating the highly processive character of this enzyme. Commercial samples of T7 exonuclease normally are endonuclease free and therefore prolonged incubation times are possible. In theory, this exonuclease can be used in combination with any restriction endonuclease because it removes almost all of the (+)strand in less than 15 min using nicked M13mpl8 as substrate. The T7 gapping protocol is given below. 9 Procedure 8: Gapping Using T7 Exonuclease. Exonuclease T7 functions in either buffer used for performing the nicking reaction. 1. Add I0 units T7 exonuclease per microgram double-stranded DNA. 2. Briefly vortex and spin down the solution. 2. Incubate at 37 ° for 30 min. 3. Heat inactivate the exonuclease at 70 ° for 15 min and place directly into a 37° heating block for 20 min. 4. Remove a 14-/A sample for gel analysis (see Fig. 2) and then place on ice. A band that runs close to a single-stranded DNA marker should be evident. Alternatively, exonuclease III can be used as the gapping enzyme. It digests double-stranded DNA containing a free 3' terminus in the 3' ~ 5' direction. We have found that exonuclease III gaps best in a buffer containing - 1 2 0 mM NaCI, 50 mM Tris-HC1 (pH 8), 6 mM MgCI2, 10 mM DTT, and 15 units of exonuclease per microgram of nicked DNA. Gel analysis of the reaction product revealed a distinct band whose electrophoretic mobility increases progressively with longer incubation time, 28 indicative of a distributive gapping mechanism, z9 Therefore, this enzyme is ideally suited for gapping in conjunction with a restriction endonuclease that produces a nick at the 3' side of the mutation in the (+)strand. We have frequently used the M13mpl8 DNA vector together with the NciI nicking/exonuclease III gapping combination to degrade the wildtype (+)strand. NciI has sites at positions 1924, 6247, 6248, and 6838 in this vector. Interestingly, neither DNA strand of the double site in the polylinker of M13mpl8 (CCCGGG 6247/8) is hydrolyzed when dCMPS is present in the (-)strand. 8 The nearest downstream NciI site to the polylinker is at position 6838. Because exonuclease III gaps nicked DNA at a 28 W. Schmidt, "Untersuchungen zum Abbau von DNA mit Exonuklease III," Diplomarbeit. University of Goettingen, Goettingen, Germany, 1986. 29 S. G. Rogers and B. Weiss, this series, Vol. 65, p. 201.
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
203
rate of about 150 nucleotides/min,9'29 a mismatch site in the middle of the polylinker in an insert of 1000 bases would require approximately 20 min (allowing a safety margin) to gap past the mismatch. Exonuclease III may of course be used in conjunction with restriction endonucleases other than NciI, provided that the buffer conditions are adjusted accordingly. Note that higher salt concentrations are essential for reproducible gapping of the DNA with exonuclease III.
Procedure 9: Gapping Using Exonuclease IIl 1. To the nicked DNA from procedure 7, add NaC1 to a concentration of 120 mM (3/zl of a 5 M solution for the AvaInicked and 6/zl for the NciI-nicked DNA) Exonuclease III, 300 units 2. Briefly vortex and spin down the solution. 3. Incubate at 37° for the time period required to gap at least several hundred bases past the mismatch. 4. Heat inactivate the exonuclease at 70° for 15 min and place directly into a 37° heating block for 20 min. 5. Remove an 8-/M sample for gel analysis (see Fig. 2) and place on ice. Repolymerization. The DNA resulting from either gapping procedure described above must be repolymerized to the double-stranded form to obtain high transformational efficiencies. Even exhaustive gapping with either T7 exonuclease or exonuclease III leaves a small stretch of doublestranded DNA that can be used as the primer for repolymerization. Obviously, the mutant DNA strand is used as a template for the reaction resulting in the formation of a mutant homoduplex.
Procedure 10: Formation of Mutant Homoduplex I. To the gapped DNA solution prepared as described above, add
E. coli DNA polymerase I (not the Klenow fragment), l0 units dNTP mix (4 x ), 5/zl ATP (10 raM), 20/zl T4 DNA ligase, 10 units 2. Incubate at 16° overnight or at 37° for 2 hr. 3. Remove a 14-/zl sample for gel analysis (see Fig. 2).
Transformation of Competent Cells As mentioned previously, special cell lines are not required for transformation when using the phosphorothioate-based mutagenesis method. We
204
MUTAGENESIS AND GENE DISRUPTION
[13]
recommend the cell lines SMH503° or TG1, 31 because they yield consistently high transformational efficiencies even if little or no RFIV DNA is visible by agarose gel electrophoresis after repolymerization. A transfection protocol is described for competent cells prepared by the CaC12 method (see procedure 11). There are a number of other transformational protocols available. 32'33 In some instances the presence of nucleoside triphosphates can result in low plaque yields. 34 This may be countered by purification of the DNA by precipitation or 35 one-step spun-column chromatography. 36
Procedure 11: Preparation of Competent Cells 1. Add 3 ml of an overnight culture (see procedure 1) to 100 ml of sterile 2YT medium in a 250-ml flask. 2. Incubate in a shaker at 37° until the A660 is 0.6 ( ~ 1 hr). 3. Transfer cells to suitable sterile centrifuge tubes, cap, and spin at -3000 g for 15 min at 4 °. 4. Discard the supernatant and resuspend the cells in a total volume of 50 ml prechilled sterile 50 mM CaCI2 solution. 5. Leave on ice for 30 min. Centrifuge as in step 3. 6. Discard the supernatant and resuspend the cells in a total volume of 20 ml prechilled sterile 50 mM CaCI2 solution. 7. The cells can be used for 1 week if stored at 4 °. 8. For long-term storage of the competent cells, take 10 ml of the competent cells and mix gently with 2 ml of 87% (v/v) glycerol (sterilized by autoclaving). 9. Portion the cells (300/~1), using a wide-bore disposable pipette, into sterile polypropylene tubes and quick freeze using liquid nitrogen. Store at - 80 °. The frozen cells can be used for several months without a serious decrease in transformational ability. 30 j. E. LeClerc, N. L. Istock, B. R. Saran, and R. Allan, J. Mol. Biol. 180, 217 (1984). 31 p. Carter, H. Bedouelle, and G. Winter, Nucleic Acids Res. 13, 4431 (1985). 32 D. Hanahan, in " D N A Cloning: A Practical Approach" (D. M. Glover, ed.), p. 109. IRL Press, Oxford, England, 1985. 33 C. T. Chung, S. L. Niemela, and R. H. Miller, Proc. Natl. Acad. Sci. U.S.A. 86, 2172 (1989). 34 A. Taketo, J. Biochem. (Tokyo) 75, 895 (1974). 35 T. Maniatis, E. F. Fritsch, and J. Sambrook, in "Molecular Cloning: A Laboratory Manual," p. 461. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982. 36 T. Maniatis, E. F. Fritsch, and J. Sambrook, in "Molecular Cloning: A Laboratory Manual," p. 466. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982.
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
205
Procedure 12: Transformation of Escherichia coli with the Mutated DNA 1. Place 5 sterile polypropylene tubes containing 3-ml portions of B-broth soft agar in a 55 ° water bath. 2. Place a tube containing 300/~1 of competent cells (procedure 11) on ice. 3. Dilute 20/zl repolymerized DNA with 30/A sterile H20. 4. Add 2-, 5-, 10-, and 20-/~1 aliquots of diluted DNA to the competent cells. 5. To the fifth tube make a mock transfection with 20 tzl sterile H20 used in diluting the DNA. Swirl the tubes gently to mix the contents and place on ice for - 3 5 min. 7. Combine 1400/zl flesh cells with 280/zl IPTG solution and 280 tzl X-Gal solution. 8. To each aliquot of transformed competent cells (from steps 4 and 5) add 270/xl of fresh cell mix from step 7. 9. Add 3 ml top agar to each tube and pour immediately onto plates prewarmed to 37 °. Allow to set and invert. 10. Incubate overnight at 37°. 11. Pick two to five plaques and prepare single-stranded DNA for direct genotypic screening of mutants by DNA sequencing. .
Principle of Phosphorothioate-Based Plasmid Mutagenesis Method As stated in the introduction, the phosphorothioate-based mutagenesis method has been extended to double-stranded or plasmid DNA vectors. 10 This is an important extension of the previous protocols because the user now avoids time-consuming subcloning steps of the gene of interest into a M13 type single-stranded DNA vector. 22 Alternatively, subcloning may also be avoided if the gene is present in a phagemid vector. 14 However, working with these constructs may require prior experience to obtain single-stranded DNA suitable for mutagenesis. The plasmid mutagenesis method is based on the creation of a specific region of single-stranded DNA to which a mismatch oligonucleotide can anneal. A strand asymmetry is created on polymerization by the specific incorporation of phosphorothioate internucleotidic linkages in a certain region of the DNA. A subsequent step using a restriction endonuclease that is unable to hydrolyze phosphorothioates containing DNA removes all the remaining wild-type DNA. Mutational efficiencies using this protocol have reached those obtained with high-efficiency methods for singlestranded DNA.
Mutatio~~HindIII Pst l
Site
A
~ HindIII/Ethidium Bromide
3"
BOQ O® cQQ ExonucleaseIII O
~
Productivelygapped
1. Add mismatch primer ~ e D 2. "1"7DNA polymerase dGTP(zS mix 3. T4 DNA Ligase Pst I
Pst I
o00
Pst l
.4--- Mutant Heteroduplex
I Pstl nicking/linearization
0 I Q
T7 Exonuclease
F
~
E . c o / i DNA Polymerase I 4 dNTPs,T4 ligase, ATP
G
Q
Transform Competent Cells
~
MutantHomoduplex
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
207
Plasmid Mutagenesis Strategy Scheme II outlines one combination of enzymatic reactions that can be used for the mutagenesis of double-stranded DNA. The first step involves the site-specific nicking of the vector in the vicinity of the desired mutation (Scheme IIB). This reaction, which is not strand specific, is accomplished by incubation of the plasmid DNA with a restriction endonuclease in the presence of ethidium bromide. The nick can be used as the starting point for digestion by a nonprocessive exonuclease such as exonuclease III (Scheme IIC). Two differently gapped products are obtained, depending on which strand contains the nick. We have designated the DNA species that has a small stretch of single-stranded DNA complementary to the synthetic mutant oligonucleotide as "productively gapped." After strand-selective hybridization of the primer to the "productively gapped" DNA, the surrounding gaps are filled in by polymerization using T7 DNA polyrnerase, three dNTPs, and one dNTPaS. As with the single-stranded mutagenesis method described above, the choice of dNTPaS is dependent on the restriction enzyme to be used in the following step (Scheme liD). As shown, the desired mutation is present as a heteroduplex in only one of the various plasmid DNA species present after polymerization. The introduction of phosphorothioate groups into the mutant strand during the polymerization step allows for strand selection, which is required for highly efficient plasmid mutagenesis. Subsequent reaction with a restriction endonuclease such as PstI (Scheme IIE), which is unable to cleave phosphorothioate-containing DNA (Table I), hydrolyzes all the wild-type DNA in solution. This includes the linearization of roughly 50% of the DNA population that was not "productively gapped." In contrast, the heteroduplex DNA containing phosphorothioates in the mutant strand at the recognition site of PsiI is only nicked. This situation is very similar to the mutagenesis procedure described above. The nick is again used as the starting point for exonuclease digestion (Scheme IIF). The mutant
SCHEME II. Schematic of the oligonucleotide-directed plasmid mutagenesis technique. (A) Plasmid DNA site of mutation and several convenient restriction endonuclease sites; (B} products from HindIII/ethidium bromide nicking reaction; (C) products of limited exonuclease III digestion; (D) the mutant heteroduplex after annealing of the mutant oligonucleotide and polymerization; (E) products from PstI nicking/linearization reaction; (F) products after T7 exonuclease digestion of the wild-type DNA; (G) repolymerized mutant homoduplex. O, Mismatch bases within the mismatch oligonucleotide. Heavy lines indicate the area where phosphorothioates have been incorporated. The plasmids that have been linearized are crossed off because they transform inefficiently.
208
MUTAGENESIS AND GENE DISRUPTION
[13]
FIG. 3. Analysis of pUC19 plasmid mutagenesis intermediates by 2% (w/v) agarose gel electrophoresis.
strand is then used as a template for repolymerization, resulting in the formation of a mutant homoduplex species carrying the desired changes in both strands (Scheme IIG). Figure 3 shows the agarose gel electrophoretic results after each stage of a typical plasmid mutagenesis experiment. The novice should compare results after each stage of the procedure with the results presented in Fig. 3.
Preparation of Plasmid DNA The purity of the plasmid DNA is very important for the successful completion of the mutagenesis protocol. There are two important variables that we have found to decrease mutational efficiencies significantly. First, the plasmid preparation should be free of large amounts of concatemeric DNA. If this is a problem, the plasmid can be grown in a RecA- strain of E. coli such as JM109. Second, small amounts of RNA that remain after many typical plasmid isolation procedures (including CsC1 centrifugation) must bc removed. In our hands plasmid DNA isolated according to the Qiagen plasmid DNA maxipreparation procedure is of sufficient purity for use without further manipulation.
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
209
Preparation of Mutant Heteroduplex
Site-Specific Restriction Endonuclease/Ethidium Bromide Nicking of Plasmid To create a stretch of single-stranded DNA to which the mutant oligonucleotide can anneal, the double-stranded DNA must be nicked in a region near the site of mutation. Although it is unimportant if the nicking site is upstream or downstream from the mutation site, it is advantageous if it is within several hundred base pairs so that subsequent digestion by the exonuclease does not have to proceed very far. Nicking can be carried out by incubation of the DNA with one of a number of restriction endonucleases that is unable to cleave both strands of DNA when incubated in the presence of ethidium bromide. 10,37-41Procedure 13 outlines the protocol for nicking plasmid DNA with the restriction enzymes HindlII and EcoRI. The use of these enzymes should be most universal because they are found at either the upstream or downstream end of a number of different multiplecloning sites in popular plasmid vectors. The protocol requires a large amount of plasmid DNA, which allows convenient monitoring of each enzymatic reaction by agarose gel electrophoresis. The amount of DNA and reaction volumes can be scaled down if desired. If one of the restriction enzymes mentioned above does not have a site, or is present at multiple positions, within the vector to be mutated then nicking conditions must be determined for another enzyme. Normally, it is a simple process to find proper nicking conditions with alternative restriction endonucleases. Almost all restriction endonucleases cleave two strands of double-stranded DNA in a stepwise fashion. It is believed that the decrease in enzyme-catalyzed hydrolysis of the second strand in the presence of ethidium bromide is due to the intercalation of the dye into the relaxed DNA. The delay between scission of the two strands can be optimized by testing several reaction conditions using suboptimal salt, pH, and/or temperature conditions for the enzyme and 10-100 tzg/ml ethidium bromide. It is advisable to use an endonuclease that does not 37 G. Dalbadie-McFarland, L. W. Cohen, A. D. Riggs, C. Morin, K. Itakura, and J. H. Richards, Proc. Natl. Acad. Sci. U.S.A. 79, 6409 (1982). 38 M. Osterlund, S. Luthman, S. V. Nilsson, and G. Magnusson, Gene 20, 121 (1982). 39 R. C. Parker, R. M. Watson, and J. Vinograd, Proc. Natl. Acad. Sci. U.S.A. 74, 851 (1977). 4o D. R. Rawlins and N. Muzyczka, J. Virol. 36, 611 (1980). 4i D. Shortle and D. Nathans, Proc. Natl. Acad. Sci. U.S.A. 75, 2170 (1978).
210
MUTAGENESIS AND GENE DISRUPTION
[13]
exhibit star activity under adverse buffer conditions. 42 It is not important to nick 100% of the plasmid DNA because any unreacted DNA will be destroyed in a later step.
Procedure 13: Nicking of Plasmid DNA 1. Add the following reagents to a 1.5-ml sterile microcentrifuge tube: Plasmid DNA (Qiagen purified), -20/.tg Buffer D (10×), 24/zl Ethidium bromide solution (500/.Lg/ml), 20/zl HindlII, 200 units 2. 3. 4. 5.
Adjust the volume to 240/xl using sterile distilled H20. Briefly vortex and spin down the solution. Incubate at 30 ° for 60 min. Remove 2/zl for agarose gel electrophoresis (see Fig. 3).
Alternatively, the DNA can be nicked using EcoRI with the conditions given below. 1. Combine the following reagents: Plasmid DNA, -20/.tg EcoRI nicking buffer (10 × ), 26/zl Ethidium bromide (500/zg/ml), 72/zl EcoRI, 600 units 2. 3. 4. 5.
Adjust the volume to 260 txl using sterile distilled H20. Briefly vortex and spin down the solution. Incubate at 16° for 15 hr. Remove 2/xl for agarose gel electrophoresis (see Fig. 3).
Successful nicking of the plasmid should result in greater than 50% nicked DNA as determined by agarose gel electrophoresis (e.g., see Fig. 3). It may be necessary to increase or decrease the amount of enzyme in the reaction. After the nicking reaction is complete the enzyme and ethidium bromide must be removed. This is accomplished by phenol extraction followed by spin dialysis using a Centricon-100 (Amicon, Danvers, MA) microconcentrator (procedure 14). 42 A l t h o u g h the restriction e n d o n u c l e a s e E c o R I is k n o w n to exhibit star activity, it has been s h o w n that w h e n Co 2+ is u s e d as the metal cofactor, the e n z y m e exhibits very stringent substrate specificity. 43 43 j. L. W o o d h e a d , N. Bhave, and A. D. B. Malcolm, Fur. J. Biochem. 115, 293 (1981).
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
211
Procedure 14: Extraction and Spin Dialysis 1. Add 200/~l buffer-equilibrated phenol and vortex vigorously for 30 see.
2. Briefly spin the tube and remove the aqueous top layer with a pipette. 3. Repeat steps 1 and 2 using 200 t*1 H20 saturated chloroform/isoamyl alcohol (24 : 1). 4. Repeat steps 1 and 2 using 1 ml H20 saturated diethyl ether. 5. Remove the final traces of ether by heating the tube with the cap open at 37° for 10 min. 6. Dilute the sample with 2 ml of distilled sterile H20. 7. Add the sample to a Centricon-100 and spin for 20 min at 1000 g with a Sorvall SS34 fixed angle rotor at room temperature. 8. Repeat steps 6 and 7 two more times. 9. Collect the sample, and transfer solvent-containing DNA (50-60 >1) to a sterile 1.5-ml microcentrifuge tube.
Preparation of Single-Stranded Gap to Which Mutant Oligonucleotide Can Anneal The DNA is now prepared for reaction using a 3' --~ 5' nonprocessive exonuclease that will gap the nicked DNA in a defined direction. The size of the gap must encompass not only the site to which the mutant oligonucleotide is to anneal, but it must also gap past the second restriction endonuclease site. A nonprocessive exonuclease (one in which the time of incubation determines the number of hydrolyzed bases) is required so that on repolymerization phosphorothioate residues are incorporated into distinct regions of the DNA. The protocol for gapping using the enzyme exonuclease III is given below. Alternatively, the 5' ~ 3' activity of T5 exonuclease can also be used for this step 44 and this enzyme will be commercially available in the near future from Amersham and United States Biochemicals.
Procedure 15: Gapping with Exonuclease III 1. Adjust the volume of the solution to 80/~1 with H20 and add Buffer C (10 × ), 10/~1 NaCI (1 M), 4/~1 Exonuclease III, 100 units 2. Briefly vortex and spin down the solution. 44 j. R. Sayers and F. Eckstein, Nucleic Acids Res. 19, 4127 (1991).
212
MUTAGENESIS AND GENE DISRUPTION
[13]
3. Incubate at 37° for the time required for digestion past the mismatch and the second restriction enzyme site. 45 4. Remove 2/zl for agarose gel electrophoresis (see Fig. 3). It is important that the products of the gapping reaction are checked by agarose gel electrophoresis. Digestion of the DNA should not proceed to such an extent that on repolymerization, phosphorothioate residues are incorporated into the wild-type strand in the vicinity of the second restriction endonuclease site. This will protect the unwanted wild-type DNA from hydrolysis, resulting in a decrease in mutational efficiency. In general, there should be a distinct increase in mobility of the gapped DNA when compared to a marker of the nicked DNA sample. Figure 3 shows an example of a T5 exonuclease-gapped pUC19 plasmid in which several hundred bases were removed.
Annealing of Mutant Oligonucleotide The oligonucleotide sequence will determine the type of mutation (point mutation, insertion, or deletion; see discussion above). Mutagenesis with plasmid DNA has two strands to which the mutant oligonucleotide can potentially anneal. Therefore, careful consideration must be given to the sequence of the oligonucleotide because it is dependent on the directionality of the exonuclease employed (procedure 15) and whether the initial nicking of the plasmid occurred upstream or downstream of the site of mutation (Scheme III). Before the oligonucleotide can be annealed to the gapped plasmid (procedure 16) it must be phosphorylated (procedure 3). One advantage of the plasmid mutagenesis procedure is that nonspecific binding of oligomer to target DNA is decreased because only a portion of the plasmid remains double-stranded after exonuclease digestion.
Procedure 16: Annealing of Mutant Primer to Single-Stranded Region of Plasmid 1. Add the following reagents to the solution containing the gapped DNA: NaCI (1 M), 10/xl Two to three molar equivalents of phosphorylated primer (procedure 3) with respect to the amount of plasmid DNA 2. Briefly vortex and spin down the solution. 45 See p r o c e d u r e 9 for information regarding the time of the gapping reaction using exon u c l e a s e III.
[13]
213
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS EcoRI
Mutation Site
Hind ll
,a,
Ethidium Mutation Site
B
Bromide
5'
Mutation Site
5'
Limited Exonudease III gapping
C
5'
5'
3'~
~
3" Anneal mutant oligomer
D 5' 3'===
~
~
5'~= 3'
oligo A A
oligo B
SCHEMEIII. Representation of two differently gapped plasmid species, both of which were "productively gapped" using exonuclease IIl. Downstream (HindIII) or upstream nicking (EcoRI) can dictate the proper sequence of the mutant oligonucleotide (either A or B) required for the production of the same genotypic change. The lettering is the same as for the steps in Scheme II. The DNA species can be nicked either upstream or downstream of the site of mutation (B). Subsequent gapping using the 3' ~ 5' activity of exonuclease III leaves two different DNA species (C). The proper mutant oligonucleotide (primer A or primer B) is required to anneal to the single-stranded DNA region.
3. I n c u b a t e at 70 ° for 10 m i n a n d t h e n p l a c e the t u b e into a 56 ° h e a t i n g b l o c k a n d cool s l o w l y to 37 ° o v e r - 3 0 min.
Polymerization Reaction T h e g a p p e d D N A is filled in u s i n g n a t i v e T7 D N A p o l y m e r a s e , t h r e e d N T P s , and one d N T P ~ S analog. The choice of nucleoside phosphorothioate is d e p e n d e n t o n the r e s t r i c t i o n e n d o n u c l e a s e u s e d in the n e x t step. W e s u g g e s t u s i n g dGTPo~S in c o m b i n a t i o n with PstI.
214
MUTAGENESIS AND GENE DISRUPTION
[13]
An improvement in the plasmid mutagenesis procedure is the use of T7 DNA polymerase for the creation of the mutant heteroduplex. This enzyme does not strand displace the mutant oligonucleotide on polymerization even when used at 37°. Products of the reaction should be checked by agarose gel electrophoresis. There should be a new DNA band that migrates differently than the supercoiled plasmid DNA marker on the gel. The polymerized DNA must be separated from the salt, nucleotides, and enzyme before being nicked.
Procedure 17: Preparation of Mutant Heteroduplex 1. Add the following reagents to the sample solution: Buffer A (10 × ), 21/xl dGTPaS mix, 20 t~l ATP (10 mM), 20 t~l T7 DNA polymerase, 10 units T4 ligase, 15 units 2. Add sterile distilled H20 to bring the volume to 210/xl. 3. Briefly vortex and spin down the solution. 4. Incubate at 37 ° for 2 hr. 5. Remove 6/.d for agarose gel electrophoresis (see Fig. 3). 6. Repeat procedure 14 to prepare the DNA for the nicking/linearization reaction. Preparation of Mutant Homoduplex
Strand-Selective Hydrolysis of Wild-Type DNA The dGMPS-containing DNA is now ready for the strand-selective nicking/linearization reaction catalyzed by PstI (procedure 18). There are a number of other restriction endonucleases that can be used for this step 8'16-19 (see Table I) as long as the enzyme chosen has its recognition site(s) located within the region protected by phosphorothioate groups in the mutant strand. As seen in Scheme II, reaction with PstI will linearize roughly 50% of the DNA in solution. Therefore, analysis by agarose gel electrophoresis should reveal a large amount of linear DNA as well as a nicked plasmid product. Most importantly, there should be no trace of covalently closed circular DNA seen on the gel.
Procedure 18: Pst! Nicking and Linearization Reaction 1. After extraction and spin dialysis, bring the volume to 85/xl with sterile HzO and add
[13]
HIGH-EFFICIENCY SITE-DIRECTED MUTAGENESIS
215
Buffer D (10 x ), 10/~1 PstI, 70 units 2. 3. 4. 5.
Briefly vortex and spin down the solution. Incubate at 37° for 80 min. Remove 3/xl for agarose gel electrophoresis (see Fig. 3). Repeat extraction and spin dialysis (procedure 14). 46
Exonuclease Digestion of Mutant Heteroduplex Wild-Type Strand The nicked DNA is gapped most efficiently using T7 exonuclease and the conditions given in procedure 19. This enzyme is very processive and will digest virtually the entire wild-type strand after short incubation periods. In many cases it is not possible to observe the fully gapped product after reaction by agarose gel electrophoresis (compare with Fig. 3) because of the poor ability of DNA to bind ethidium bromide (because it is essentially single-stranded DNA). It is possible, however, to see that the nicked and linear DNA resulting from procedure 19 has been digested when compared to the DNA analyzed after PstI nicking. In some instances, where the nick is downstream to the mutation in the wild-type strand, the user may use exonuclease III for this step. 10However, we have obtained the most reproducible results using T7 exonuclease.
Procedure 19: Gapping of Nicked Mutant Heteroduplex 1. Adjust the volume of the solution to 90/xl with H20 and add Buffer C (10 x ), 10/zl T7 exonuclease, 100 units 2. Briefly vortex and spin down the solution. 3. Incubate at 37° for 30 min. 4. Heat inactivate the enzyme by incubation at 70° for 10 min and place directly into a 37° heating block for 20 min. 5. Remove 10/.d for agarose gel electrophoresis (see Fig. 3). In theory, the gapped DNA could be used directly to trasnsform competent cells because the wild-type DNA has been almost completely hydrolyzed. However, we have observed an increase in mutational efficiency of up to 20% if the DNA is first repolymerized with E. coli DNA polymerase I, 46 It is important to repeat the extraction/dialysis procedure after PstI nicking because this enzyme binds tightly to the nicked phosphorothioate-containing DNA, which can inhibit exonuclease digestion?
216
MUTAGENESIS AND GENE DISRUPTION
[13]
TABLE II TROUBLE-SHOOTING FOR MUTAGENESISOF SINGLE-AND DOUBLE-STRANDED DNA VECTORS Problem
Possible causes
Polymerization results in RFII DNA
Nonspecific priming by mutant oligonucleotide Inactive ligase
Incomplete nicking by the restriction endonuclease Restriction endonuclease reaction results in linearization of the DNA
Restriction endonuclease activity too low
Incomplete nicking of plasmid
Restriction endonuclease activity too low
No cccDNA observable by agarose gel electrophoresis after repolymerization
Low product yield DNA was destroyed during exonuclease reaction
False restriction endonuclease/ dNTPaS combination Restriction enzyme site in mutant oligonucleotide sequence
Remedy Decrease primer concentration Check enzyme and ATP concentrations Increase incubation time or enzyme concentration Check Table I for correct dNTPc~S/restriction enzyme combination Choose another restriction enzyme or protect site in the oligonucleotide with phosphorothioatecontaining oligomer (see text) Increase enzyme concentration or reaction temperature Attempt transformation Repeat procedure using fresh gapping buffer and new batch of T7 exonuclease
f o u r n o r m a l d N T P s , A T P , a n d T4 ligase. 10 T h e n i c k - t r a n s l a t i o n a c t i v i t y o f t h e p o l y m e r a s e m i g h t r e m o v e s o m e D N A still b o u n d to t h e m u t a n t o l i g o n u c l e o t i d e a f t e r e x o n u c l e a s e d i g e s t i o n , w h i c h c o u l d a c c o u n t for t h e i n c r e a s e in m u t a t i o n a l e f f i c i e n c i e s .
Procedure 20: Preparation of Mutant Homoduplex 1. A d d t h e f o l l o w i n g r e a g e n t s to t h e s o l u t i o n c o n t a i n i n g the g a p p e d DNA: B u f f e r A (10 x ), 5 ~1 D N A p o l y m e r a s e I, 10 u n i t s 4 x dNTP Mix, I0/zl A T P (10 m M ) , 2 0 / x l T4 D N A l i g a s e , 15 units 2. A d d s t e r i l e d i s t i l l e d H 2 0 to b r i n g t h e v o l u m e to 220/~I. 3. Briefly v o r t e x a n d s p i n d o w n t h e s o l u t i o n .
[ 13]
HIGH-EFFICIENCY SITE-DIRECTEDMUTAGENESIS
217
4. Incubate at 37° for 3 hr. 5. Remove 14 p,l for agarose gel analysis (see Fig. 3). 6. Remove 2 tzl for transformation of competent cells (procedure 21). Transformation of Competent Cells It is sometimes difficult to observe a band corresponding to covalently closed circular plasmid DNA. We recommend that a sample of the DNA be transformed nonetheless. The competent cells prepared according to procedure 11 can be used for the transformation with the mutated DNA according to procedure 21. Procedure 21: Transformation o f Competent Cells
1. Place tube containing 100 /zl of competent cells, prepared as in procedure 11, on ice. 2. Add at least 2/zl of repolymerized DNA from procedure 20, mix gently, and place on ice for 10-40 min. 3. Add 300/zl 2YT medium and shake at 37° for 60 min. 4. Take 2, 10, and 80/zl and spread onto agar plates containing the appropriate antibiotic selection marker. 5. On another plate spread 10/zl of competent cells that have not come in contact with any DNA. 6. Incubate the plates overnight at 37° and pick two to five colonies for DNA characterization. Troubleshooting One advantage of performing in vitro site-directed mutagenesis is that each step of the procedure may be conveniently analyzed by agarose gel electrophoresis. We recommend that the novice check the products after each enzymatic reaction before proceeding to the next step. Figures 2 and 3 give examples of how the different DNA species should appear after agarose gel analysis. It is important to run the gels in the presence of ethidium bromide and 2-mercaptoethanol because these compounds increase resolution as well as the stability of the DNA. 7'47 Table II gives a brief summary of potential pitfalls that might become evident after electrophoretic analysis. Acknowledgments We kindlyexpress our appreciationto A. Fahrenholzand A. Kroggelfor expert technical assistance and to R. Mackinfor critical reading of the manuscript. 47B. V. L. Potter and F. Eckstein,J. Biol. Chem. 259~ 14243 (1984).
218
MUTAGENESISAND GENEDISRUPTION
[14]
[14] P o l y m e r a s e C h a i n R e a c t i o n - B a s e d Point Mutagenesis Protocol
By L.-J.
ZHAO,
Q. X.
ZHANG,
and R. PADMANABHAN
Introduction Since its inception in 1985, the polymerase chain reaction I (PCR) has become an extremely useful technique in molecular biology. Its use ranges from clinical diagnostics to structure and functional studies involving sitedirected mutagenesis. The PCR involves the selective amplification of a segment of DNA template flanked by two synthetic complementary oligodeoxynucleotide primers by repeated cycles of three basic steps: heat denaturation of the template DNA, annealing of the primers to the template to form stable duplexes, and extension of the 3' ends of the primers by a DNA polymerase. Each new cycle of PCR amplification gives rise to twice the number of copies of the template DNA from the previous cycle as a result of annealing and extension of these primers by DNA polymerase. Thus the region between the two primers is amplified exponentially by approximately 2n-fold, where n is the number of cycles. 1,2Initially, Escherichia coli DNA polymerase I (Klenow fragment) was used for DNA synthesis from the annealed primers,3 but was replaced later by the thermostable DNA polymerase from Thermus aquaticus (Taq), which allowed the automation of this amplification method. One of the powerful applications of the PCR is in site-directed mutagenesis protocol to carry out the structural and functional analysis of genes encoding proteins or regulatory elements. Deletions, insertions, and point mutagenesis can be carried out using the PCR. Previous methods for deletion mutagenesis involve the wide use of either BAL 31 nuclease, which progressively shortens a double-stranded DNA fragment from both the 5' and 3' ends, or exonuclease III, which digests the target DNA from the 3' ends. The latter method can be applied to digest target DNA unidirectionally. 4 The extent of digestion in both cases is controlled by incubation time or the temperature of the reaction or both. Before the i R. K. Saiki, D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Erlich, Science 239, 487 (1988). 2 K. B. Mullis and F. A. Faloona, this series, Vol. 155, p. 335. 3 R. K. Saiki, S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. A r n h e i m , Science 230, 1350 (1985). 4 S. Henikoff, Gene 28, 351 (1984).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993by AcademicPress, Inc. All rights of reproduction in any form reserved.
[14]
PCR-BASED POINT MUTAGENESIS PROTOCOL
219
advent of the PCR technique, the insertion mutagenesis method commonly involved the ligation of synthetic restriction site linkers to the target DNA linearized by partial digestion with nonspecific deoxyribonuclease I (DNase I) in the presence of Mn 2÷ . An alternate method for obtaining linear DNA is by digestion with a restriction enzyme having a tetranucleotide recognition site in the presence of ethidium bromide. 5,6The insertional mutagenesis is random when DNase I is used, or limited to those regions containing the restriction sites on the target DNA. Several methods have been described to introduce point mutations along the segment of DNA. 6 These methods include mutagenesis by (1) treatment with sodium bisulfite, which deaminates deoxycytidine to deoxyuridine, resulting in the substitution of an A : T base pair for a G : C base pair in approximately 50% of the template molecules after one round of replication, 7'8 (2) enzymatic incorporation of nucleotide analogs,9 or misincorporation of normal nucleotides or ct-thionucleotide l°'~t by DNA polymerases. Oligo-directed mutagenesis has been extremely useful to generate deletion, insertion, and point mutations at preselected sites on the DNA molecule. Prior to mutagenesis, the target DNA is cloned into an M13 vector so that singlestranded wild-type DNA template can be produced. The oligo mutagen is then annealed to this template, producing a noncomplementary (looped out) region on the oligo primer or on the template, resulting in an insertion or a deletion, respectively. A third possibility is a base pair mismatch between the template and the primer in the case of point mutagenesis. Although these methods are efficient, they are still time consuming because of the number of steps involved, such as cloning of the target DNA into an MI3 vector, screening for the mutants by DNA sequence analysis, and recloning of the mutant DNA segments back into the parent plasmid for functional studies. On the other hand, the PCR-based mutagenesis methods 12-16 are simple and rapid compared to the more conventional methods described above.
5 R. C. Parker, R. M. Watson, and J. Vinograd, Proc. Natl. Acad. Sci. U.S.A. 74, 851 (1977). M. Smith, Annu. Rev. Genet. 19, 423 (1985). 7 D. Shortle and D. Nathans, Proc, Natl. Acad. Sci. U.S.A. 75, 2170 (1978). 8 D. Botstein and D. Shortle, Science 229, 1193 (1985). 9 W. Mueller, H. Weber, F. Meyer, and C. Weissmann, J. Mol. Biol. 124, 343 (1978). to R. A. Zakour and L. A. Loeb, Nature (London) 295, 708 (1982). tt D. Shortle, P. Grisafi, S. Benkovic, and D. Botstein, Proc. Natl. Acad. Sci. U.S.A. 79, 1588 (1982). t2 R. Higuchi, B. Krummel, and R. K. Saiki, Nucleic Acids Res, 16, 7351 (1988). t3 F. Vallette, E. Mege, A. Reis, and M. Adesnik, Nucleic Acids Res. 17, 723 (1989). 14 H. Kadowaki, T. Kadowaki, F. E. Wondisford, and S. 1. Taylor, Gene 76, 161 (1989).
220
MUTAGENESIS AND GENE DISRUPTION
[14]
We have utilized this PCR-mediated point mutagenesis protocol to analyze the function of the nuclear localization signal (NLS) domain in the 140-kDa adenovirus DNA (AdPol) that is required for replication of the adenovirus genome. AdPol contains three clusters of basic amino acids (R8ARR, R25RRVR, and R41ARRRR, BS I-III) which resemble the NLS domain of the simian virus 40 (SV40) large T antigen type, and therefore could potentially be involved in the nuclear targeting of AdPol or the heterologous cytoplasmic protein, E. coli fl-galactosidase (fl-Gal). 17 To examine the contribution of the basic amino acids in each cluster and the interactions between these positively charged regions in the nuclear targeting function, the PCR point mutagenesis protocol described here was used to introduce mutations into BS I-BS III (Fig. 1). Principle of Method To analyze the function of a DNA segment, either encoding a protein or a regulatory region, saturation mutagenesis in a defined region is a useful experimental approach to examine the contribution of the defined region to the overall function. It is also important that point mutations could be introduced at will along any stretch of DNA without the necessity of having convenient restriction sites in close proximity to the mutagenesis site. In the methods described to date for point mutagenesis by the PCR, the desired mutation is incorporated into one of the oligo primers. Amplification between the normal and mutant primers gives rise to the desired mutant DNA fragment. One drawback of this approach is that it requires one mutant oligo for each desired point mutation. Moreover, in cases in which the mutational analysis of a region of DNA is exploratory due to a lack of information regarding its function, it is best to produce several nucleotide substitutions at defined regions in the target DNA in a minimum number of steps. In the PCR mutagenesis protocol described here, mixedsite oligo in which the mutations are preselected along the length of a mutagenic oligo primer are chemically synthesized and used for the PCR to give rise to a mixture of PCR-generated mutant DNA fragments. To carry out functional analysis of mutations, the mutant fragment needs to be substituted for the wild-type fragment in the plasmid. To achieve this aim, the double-stranded mutant DNA fragments obtained in the first PCR were used as primers for amplification on wild-type templates. Different 15 M. Kammann, J. Laufs, J. J. Schell, and B. Gronenborn, Nucleic Acids Res. 17, 5404 (1989). 16 S. N. Ho, D. Hunt, R. M. Horton, J. K. Pullen, and L. R. Pease, Gene 77, 51 (1989). 17 C. Dingwall and R. A. Laskey, Annu. Rev. Cell Biol. 2, 367 (1986).
[14]
221
PCR-BASED POINT MUTAGENESIS PROTOCOL
A oligo
#i BS M
A
L
V
Q
A
H
R
A
I R
R
L
H
A
E
CCATG-GCC-CTT-GTT-CAA-GCT-CAC-CGG-GCC-CGT-CGT-CTT-CAC-GCA-GAG-GCG GGTAC-CGG-GAA-CAA-GTT-CGA-GTG-GCC-CGG-GCA GCA GAA-GTG-CGT-CTC oligo
A
CGC
13
GCC-CGG-GCA-GCA-GAA-GT T T BS P
D
S
G
D
Q
P
P
R
R
II R
V
R
Q
Q
P
P
R
A
CCA-GAT-TCA-GGA-GAT-CAA-CCG-CCG-CGT-CGT-CGC-GTT-CGC-CAG-CAA-CCT-CCG-CGC-GCA GGT-CTA-AGT-CCT-CTA-GTT-GGC-GGC-GCA-GCA-GCG-CAA-GCG-GTC-GTT-GGA-GGC-GCG-CGT GC-GGC-GCA-GCA-GCG-CAA T T T Oligo BS A
P
A
P
A
R
A
R
GC
14
III R
R
R
A
P
A
P
S
P
GCA-CCA-GCT-CCT-GCC-CGC-GCG-CGG-CGC-CGA-CGT-GCC-CCT-GCC-CCC-TCT-CCC CGT-GGT-CGA-GGA-CGG-GCG-CGC-GCC-GCG-GCT-GCA-CGG-GGA-CGG-GGG-AGA-GGG GG-GCG-CGC-GCC-GCG-GCT-GCA-CGG-GGA-CGG-GGG-AGA~GGG A A T A T A A A
Oligo
15
B #13 Ncol
#14
I-En--I ----
#5
~
[
# 15
# 6 ~,
BSIH
]
PCR ~mplate
Kpnl
I
] I st PCR products :# 8¢ - - - - - . ~
I
-]
I I
2nd PCR products
I Cloning and sequenceanalysis
KpnI
FIG. 1. (A and B) Two-step PCR point mutagenesis strategy. The application of this mutagenesis method is to analyze the function of three clusters of basic amino acids (BS I-III) in the N-terminal region of adenovirus DNA polymerase. 2° (A) The first-step PCR was carried out between oligo 1 as primer 1, and oligo 13, 14, or 15 as primer 2 for mutagenesis of BS I, BS II, or BS III, respectively. (B) The second PCR was carried out initially by using the first PCR product as primers, and later by the addition of oligo 1 as primer 1, and oligo 5 or 6 as primer 2. The plasmid template (pGZ1) used for PCR contained the full-length coding sequence of adenovirus DNA polymerase.
222
MUTAGENESIS AND GENE DISRUPTION
[14]
primers were then added for the second PCR step. These primers contained sequences flanking two convenient restriction sites on the wildtype plasmid template (Fig. 1B). It was found that this two-step PCR mutagenesis method was effective in introducing multiple mutations in a defined region preselected by mixed-site mutagenic oligodeoxynucleotides. Materials and Reagents The reagents for PCR were purchased from Perkin-Elmer-Cetus (Norwalk, CT) as a kit. Escherichia coil (HB101 or DH5 strain) was used for all plasmid transformations. Triton X-100 was from Sigma Chemical (St. Louis, MO). The deoxynucleoside triphosphates used initially were from the PCR kit; in later experiments they were purchased from Pharmacia (Piscataway, N J). Oligo primers were synthesized using BioSearch model 8600 (Milligen,Burlington,MA). Reagents for the synthesis of DNAprimers were purchased as kits from MilliGen Biosearch. Mixed-site oligo primers were synthesized using the automatic mixed-site capability of the DNA synthesizer. During the coupling step, each monomer is alternately sampled and mixed before arriving at the column. The PCR was carried out using the thermocycler from Coy Laboratory Products (Ann Arbor, MI) and in later experiments using the GeneAmp instrument (Perkin-ElmerCetus). Screening the mutants was done by DNA sequencing of the recombinants using the DNA isolated from 4-ml cultures. 18The DNA sequencing was carried out by the dideoxy chain termination method of Sanger, using T7 DNA polymerase (Sequenase; U.S. Biochemicals, Cleveland, OH) from a kit.
Polymerase Chain Reaction Mutagenesis Method First Polymerase Chain Reaction Step. This mutagenesis strategy involves a two-step PCR amplification, cloning, and screening the mutants by DNA sequence analysis. The first-step PCR was initiated between wildtype oligo 1,5'-CGCCATTTGACCATTCACCA-3', corresponding to nucleotides 20-40 in pRSV-LTR vector, 19and either mixed-site oligo 13 (for BS I), oligo 14 (for BS II), or oligo 15 (for BS III) (see Fig. 1) to give rise to mutant PCR products that were amplified on the wild-type PCR template 18 T. Maniatis, E. F. Fritsch, and J. Sambrook, "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982. 19 C. Gorman, G. Merlino, M. Willingham, I. Pastan, and B. H. Howard, Proc. Natl. Acad. Sci. U.S.A. 77, 313 (1980).
[14]
PCR-BASED POINT MUTAGENESIS PROTOCOL
223
(the plasmid pGZ1, containing the AdPol coding sequence cloned into the RSV-LTR plasmid vector). The PCRs were carried out using Taq polymerase (Perkin-Elmer-Cetus) or Vent DNA polymerase (New England BioLabs, Beverly, MA) according to the conditions of the manufacturer. Briefly, the incubation mixture (I00 ~1) contained 6.6 fmol of template (pGZ1 plasmid) 1 ~M each of the PCR primers, 200/~M dNTPs, 50 mM KCI, 10 mM Tris-HCl (pH 8.3 at room temperature), 1.5 mM MgCI2, 0.01% (w/v) gelatin, and 2.5 units of Taq polymerase. The sample was then overlaid with 100 ~1 paraffin oil (Fisher Scientific, Pittsburgh, PA). The PCR was carried out either in the thermocycler from Coy Laboratory Products or in GeneAmp from Perkin-Elmer-Cetus. The conditions of the three basic PCR steps were denaturation at 94° for 2 min 30 sec in the first cycle and 1 min 30 sec for cycles 2-25, annealing at 45-50 ° for 2 rain 30 sec, and polymerization at 72 ° for 2 rain. At the end of 25 cycles, the time of polymerization was extended to 10 rain, and the PCR DNA was subsequently stored at 4° until further use. The products were analyzed by electrophoresis on a polyacrylamide gel (8%) (Fig. 2A). Second Polymerase Chain Reaction Step. The first, double-stranded PCR product was diluted 10- to 50-fold, depending on the yield of the first PCR product, and was used in the second PCR as primers for elongation and subsequent amplification directed by oligo primers. These oligo primers were chosen from the sequences flanking two convenient restriction sites so that the amplified products from the second PCR could be cloned into those sites on a suitable plasmid vector and screened by DNA sequence analysis. The second PCR reaction mixture (100/zl) contained the template (1.6 fmol), 2.5-10% of the first PCR incubation mixture containing an about 500-fold molar excess of the first PCR product over the wildtype template concentration, and the rest of the components of the first PCR except the oligo primers. The DNA fragment from the first PCR product that contained the desired mutation was allowed to extend on wild-type template in the second PCR step. The second PCR was carried out for seven cycles in the absence of any additional primers except those oligos present as a carryover from the first PCR. The oligo primers 1 and 6 or 1 and 5 (Fig. 1B) were then added to the reaction at 1/zM concentration, and the PCR was continued for an additional 18 cycles. The conditions of this two-step PCR mutagenesis are given in Table I. The desired DNA fragments (311 bp between oligos 1 and 6, and 173 bp between oligos I and 5, as shown in Fig. 2B) were purified from the second PCR on an 8% (w/v) polyacrylamide gel by standard techniques of electroelution, 1~ NENSorb (Du Pont, Wilmington, DE) column purification, followed by lyophilization of the column eluate according to the protocol of the manufacturer. The purified DNA was then either digested with NcoI alone
224
MUTAGENESIS AND GENE DISRUPTION
A
1
[14]
B 2
3
1
2
3
a
b C
d e ¸
f
a
b " 1 3 9 bp c d
,311
bp
, 9 2 lap e f ,9192 bp
FIG. 2. Analysis of PCR products by electrophoresis on polyacrylamide gels. The PCR products were analyzed on polyacrylamide gels (8%) as described/8 (A) Lane 1, DNA size markers; a - f represent 344, 298,220, 201,154, and 134 bp. Lanes 2 and 3, first PCR products obtained in experiments 1 and 2 of Table I, respectively. The arrowheads indicate the 92and 139-bp products. (B) Lane l, same as in (A). Lanes 2 and 3, second PCR was initiated by the first PCR products as primers on the template pGZ1 as shown in Table I, and later continued by the addition of oligo 1 and 6 as terminal primers. The arrowheads show the final product of 311 bp (lanes 2 and 3) and, in addition, some of the first PCR product (92 bp in lane 2).
(for 173 bp) and cloned in the NcoI-SmaI-digested plasmid pLZ401,20 or digested with NcoI plus KpnI (311 bp) and cloned in the NcoI-KpnI sites of a pBR322 derivative. The transformants were screened by DNA sequence analysis. The frequency and the number of point mutations introduced into the various domains, BS I-III, were obtained from the DNA sequence analysis data, and are shown in Table II. 2o L.-J. Zhao and R. Padmanabhan, Cell 55, 1005 (1988).
[14]
PCR-BASED POINT MUTAGENESIS PROTOCOL
225
TABLE I EXPERIMENTAL CONDITIONS FOR Two-STEP POLYMERASE CHAIN REACTION MUTAGENESIS PROTOCOLa
Experiment
First PCR primers
Size of first PCR product b (bp)
Second PCR primers 1-7 cycles b 8-25 cycles
1
Oligo 1 + oligo 13
92
92 bp (50 ng)
2
oligo I + oligo 14
139
139bp (50 ng)
3
oligo 1 + oligo 15
209
209 bp (50 ng)
oligo 1 + oligo 5 (or 6) oligo 1 + oligo 5 (or 6) oligo 1 + oligo 6
Size of second PCR product' (bp) 173; 311
173; 311
311
The XhoI-linearized wild-type plasmid containing the AdPol coding sequence was used as the template (6.6 fmol for the first PCR and 1.6 fmol for the second PCR). All oligo primers were used at 1 p.M final concentration. b The primers for the second PCR during the first seven cycles were the 92-, 139-, and 209-bp double-stranded first PCR products obtained in experiments 1-3, respectively. c The sizes of the second PCR products were 173 and 311 bp, respectively, when oligo pairs 1 + 5 and 1 + 6 were used during cycles 8-25.
Discussion The PCR mutagenesis protocol described above gives rise to multiple point mutations as shown in Table II. The total number of sites varied in a mutagenic oligo dictated the number of point mutations introduced in the final product obtained as the predominant species. For example, in the mutagenesis of BS I, two sites in oligo 13 were varied, and hence 85% of the clones screened contained two point mutations (Table II). In the case of BS II and BS III, because there are three or more mixed sites in oligos 14 and 15 (Fig. 1A), there is a preponderance of clones with three or more point mutations. It is possible that during the synthesis of mixed-site oligos, the oligos with multiple mutations were disproportionately abundant over oligos with single point mutations. The nature of the nucleotide chosen for a particular site of mutation and the overall yield of the chemical synthesis at each of the mixed sites could influence the abundance of a mutant oligo in the mixture. An alternate explanation for the generation of clones with more than one point mutation is as follows. At the annealing temperature of the PCR (45-50°), the mutagenic oligos with double or more mutations could form more stable hybrids with the mutant amplified DNA, produced within a few cycles of the first PCR containing a single shared mutation, than with the wild-type template. Further amplification
226
[14]
MUTAGENESIS AND GENE DISRUPTION
T A B L E II FREQUENCY AND NUMBER OF POINT MUTATIONS ACCUMULATED USING TwO-STEP POLYMERASE CHAIN REACTION MUTAGENESlS METHOD N u m b e r of point mutations b Size o f second P C R product cloned (bp) 173 173 311 311 311
Mutant domain BS BS BS BS BS
I II I II III
F r e q u e n c y of mutation a
1
2
3
4
1.0 1.0 0.9 0.65 0.69
2 0 3 0 0
11 1 6 3 2
-10 -8 13
---3
F r e q u e n c y of mutation r e p r e s e n t s the ratio of the n u m b e r of m u t a n t s obtained divided by the total n u m b e r of t r a n s f o r m a n t s screened. b Total n u m b e r o f m u t a n t t r a n s f o r m a n t s obtained containing one, two, three, or four point m u t a t i o n s in the various domains.
resulting from this stable annealing would give rise to mutant DNA with multiple mutations in subsequent cycles. If this explanation is correct, then it would suggest that the use of excess wild-type template (in the range of 80-160 fmol) and a fewer number of cycles (about 10) in the first PCR might give rise to clones with single point mutations. However, in some structure-function studies the mutagenesis protocol described here should be useful. For example, in functional studies involving a regulatory region isolation and characterization of multiple mutations in several defined regions would be helpful to identify the contributions of individual regions toward the overall function. We also noticed that Vent DNA polymerase (isolated from the extreme thermophile Thermococcus litoralis, and sold by New England BioLabs) could also be used for this mutagenesis procedure successfully (BS I and BS II mutations in the first two rows of Table II), except that a lower annealing temperature (42-45 °) was required. At a higher annealing temperature, the frequency of mutations was considerably reduced. This attribute of Vent DNA polymerase might be related to its exceedingly stable 3' ~ 5' proofreading exonuclease activity as advertised by the manufacturer, which might effectively reduce the concentration of the mutagenic oligo, but not the fraction of the wild-type oligo annealing to the wild-type template in the first PCR. For the second PCR, the double-stranded mutant DNA obtained in the first PCR was used as primers for elongation by Taq DNA polymerase on a wild-type template. We varied the amount of the first PCR product added to the second PCR from 2.5 to 10% of the incubation mixture without the
[14]
PCR-BASED POINT MUTAGENESIS PROTOCOL
227
need for phenol extraction or gel purification step. The amount of the first PCR product needed as primers for a successful second PCR had to be established empirically. Therefore, it is recommended that the second PCR be carried out using two or three different concentrations of the first PCR product as primers for elongation and subsequent amplification. Use of double-stranded mutant DNA as primers for the PCR has been reported. 2~'z2 In one report, 2~ the two independent point mutants of exon 4 (182 bp) of the tyrosinase gene were amplified from the M13/exon 4 subclones, using oligo primers flanking the mutational sites. The doublestranded amplified DNA product was used as a primer in the second-step PCR for obtaining a longer fragment containing exons 2, 3, and 4, which was then inserted to replace the corresponding portion of the wild-type gene in an expression plasmid. In the second report, 22the mutagenic oligo containing a single mutation was used as one PCR primer. The hybrid primer, in which the 3' end contained sequences from the template region to be amplified, and the 5' end contained unique sequences not present in the template, was used as the second primer. The double-stranded mutant DNA was used as primer in the second-step PCR, in addition to two oligos, one from a site upstream of the mutation, and the other containing the 5' end sequences of the hybrid primer used in the first PCR. This strategy-'" allowed the selection of mutant DNA against the wild-type DNA. However, it required three oligonucleotides flanking a mutational site, and one mutagenic oligo for each desired point mutation. The method described here is well suited for introducing multiple mutations in a defined region. Although the transformants need to be screened by DNA sequence analysis in this procedure, different mutants can be identified by a single sequencing gel. Sequencing the transformants also achieves the purpose of verifying the accuracy of the PCR amplification, as would be required for any mutagenesis protocol. 2E L. B. Giebel and R. A. Spritz, Nucleic Acids Res. 18, 4947 (1990). 22 R. M. Nelson and G. L. Long, Anal. Biochem. 180, 147 (1989).
228
MUTAGENESIS AND GENE DISRUPTION
[15]
[15] L i b r a r i e s o f P e p t i d e s a n d P r o t e i n s D i s p l a y e d o n Filamentous Phage
By GEORGE P. SMITH and JAM1E K. SCOTT Introduction A "fusion phage" is a flamentous virion displaying on its surface a foreign peptide fused to a coat protein, and harboring the gene for the fusion protein within its genome.l'2 In this chapter we will emphasize an application for which these surface expression vectors are particularly well suited: construction of epitope libraries. In such a library--at least the kind so far constructed--the phages display "random" foreign peptides encoded by degenerate synthetic oligonucleotides spliced into the coat protein gene, the library as a whole representing up to billions of peptide sequences. 3-5 If a phage displays a peptide that is a strong ligand for an antibody or other binding protein, it can be readily affinity purified out of a library--even one containing a vast excess of nonbinding clones. Affinitypurified phages are eluted without destroying their infectivity; and the peptide sequences responsible for binding are easily ascertained by infecting the eluted phages into bacteria, propagating the resulting phage clones, and sequencing the relevant part of their viral DNAs. In this way, billions of peptide epitopes can be encompassed in a few microliters of solution and effectively surveyed for tight binding to a given protein, using simple microbiological and recombinant DNA procedures. The number of peptides that can be accommodated with this technology exceeds by at least a factor of 100-1000 the number that can be screened with conventional expression systems, in which the epitope is not displayed as part of the propagatable unit that encodes it. At the end of the chapter we will touch on more complex libraries, in which the displayed ligands are whole folded domains. Such constructs include most notably libraries of "phage antibodies" that would (it is hoped) display an array of binding specificities large enough to accommodate almost any possible antigen; these libraries hold out the promise of l G. P. Smith, Science 228, 1315 (1985). 2 S. F. P a r m l e y and G. P. Smith, Gene 73, 305 (1988). 3 j. K. Scott and G. P. Smith, Science 249, 386 (1990). 4 j. j. Devlin, L. C. Panganiban, and P. E. Devlin, Science 249, 404 (1990). 5 S. E. Cwirla, E. A. Peters, R. W. Barrett, and W. J. Dower, Proc. Natl. Acad. Sci. U.S.A. 87, 6378 (1990).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993by AcademicPress, Inc. All rights of reproduction in any form reserved.
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
229
dramatically simplifying the isolation of monoclonal antibodies to almost any antigen. To simplify the discussion we will consistently use the term ligate to refer to the molecule that specifically binds a phage-borne peptide or protein domain and ligand to refer to a molecule (whether phage-borne or not) that specifically binds a ligate. In the case of epitope libraries displaying short peptides, the ligate is almost always a binding protein such as an antibody or receptor; in this case our terminology conforms with the convention that refers to the smaller, less structured partner of a binding pair as the ligand. As the phage-borne ligand becomes more and more structured, however, especially in the case of phage antibodies, smaller and smaller molecules can serve as ligates, including simple, nonproteinaceous determinants; in such circumstances, our terminology reverses the usual convention. In this chapter we describe the basic methods of fusion-phage technology, emphasizing the random hexapeptide epitope library that has been studied in our laboratory, but touching on alternative developments.
Design of Fusion Phage Vectors Representative published fusion-phage constructs are listed in Table l, along with key features that will be described in this section; some of the entries are specific constructs rather than general cloning vectors. All are based on the Ff class of filamentous phage: the class that infects bacteria harboring the F episome and that includes wild-type strains M13, fl, and fd. The filamentous phage infection cycle 6'7 is initiated by the attachment of phage coat protein pIII (product of phage gene III) to the tip of the F pilus, followed by internalization of the single-stranded viral DNA (ssDNA). This so-called plus strand serves as template for minus-strand synthesis, which starts at a specific origin and results in a double-stranded replicative form (RF). The RF is the template for mRNA transcription, RF replication, and production of progeny ssDNA. Progeny virions are assembled, not in the cytoplasm, but rather by extrusion of ssDNA through the bacterial envelope without killing the cell or preventing cell division. As it emerges from the cell, the ssDNA acquires its extracellular sheath of coat proteins from the membrane. The coat consists of a tubular array 6 G. P. Smith, in "Vectors: A Survey of Molecular Cloning Vectors and Their Uses" (R. L. Rodriquez and D. T. Denhardt, eds.), p. 61. Butterworths, Boston, 1988. 7 p. Model and M. Russel, in "The Bacteriophages" (R. Calendar, ed.), Vol. 2, p. 375. Plenum, New York, 1988.
230
MUTAGENESIS AND GENE DISRUPTION
[15]
TABLE I FUSION-PHAGE CONSTRUCTS
Name
Gene
fUSE5
III
fAFF1 M 13LP67 phGH-M 13glII
1II 111 III
pSEX
Ili
pCBAK8
VIII
fdH
VIII
pKfdH
VIII
Features Defective for (-)strand synthesis Noncomplementary, non-selfcomplementary termini Tetracycline resistance Infectivity requires insert Same as fUSE5 Ampicillin resistance Phagemid Human growth hormone fused directly to C-terminal domain of pill Must be complemented with helper to supply wild-type plII, other phage functions Ampicillin resistance Phagemid Single-chain antibody Fv domain fused to full-length plII Must be complemented with helper to supply other phage functions Ampicillin resistance Phagemid Antibody Fab domain fused to pVIll Must be complemented with helper to supply wild-type pVIII, other phage functions Ampicillin resistance Tolerates only foreign peptides with ---six amino acids Plasmid, not a phagemid: not suitable for library construction Must be complemented with helper to supply wild-type pVIII, other phage functions Ampicillin resistance
Sequence at amino terminus of mature hybrid protein ~
Ref.
ADc~Xd3GAETVESCLAK--
b
X.chCLAK--
c
X,,flAETVESCLAK--
d
e
f
g
AEVX.NDP--
h
AEVXnNDP- -
h
" Only general cloning vectors are entered, a, V, A, D, E, or G;/3, S, P, T, or A; cb, any amino acid but W, Q, M, K, or E. b j. K. Scott and G, P. Smith, Science 249, 386 (1990).
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
231
of thousands of pVIII molecules (product of phage gene VIII) and four minor coat proteins, including five copies of pIII incorporated into the trailing tip of the emerging virion. Both pIII and pVIII are synthesized with posttranslationally cleaved signal peptides, and before incorporation into the virion are anchored in the inner membrane with the N-terminal portion (the bulk of the protein in the case ofpIII) exposed in the periplasm. The gene III protein appears to have two functional domains, each roughly 200 residues long: an exposed N-terminal domain that binds the F pilus but is not required for virion assembly, and a C-terminal domain that is buried in the particle and is an integral part of the capsid structure. In the virion, the C-terminal portion of pVIII appears to be inside the virion, close to the DNA, while the N terminus is exposed to the solvent. In most fusion-phage constructs the foreign amino acids are inserted just downstream of the pIII signal peptide, and propagation of the recombinant phage requires that the recombinant pIII retain its functions. In a wild-type filamentous phage background, defects in gene III (and in most other phage genes) cause the phages to kill the cell with almost no phage production. In mutants with a defective minus-strand origin of replication, however, gene III defects are tolerated, probably as a result of reduced RF copy number. Such phages still replicate as a result of adventitious minus-strand initiations; and if all the other phage genes are functional the virions are infective, even giving small plaques. Many of the fusion-phage vectors, including the fUSE vectors and fAFF1, derive from fd-tet, which has a tetracycline (Tc) resistance determinant spliced into the minus-strand origin. This phage can be propagated like any Tc-resistance plasmid, independently of pIII function. 8 This allows a gene III frameshift to be engineered into the cloning site, so that vector without insert is noninfective; in such "frame-shifted" vectors, only clones bearing frame-restoring inserts contribute infectious particles to a library. Use of low copy number vectors may also help accommodate inserts that partially debilitate pIII 8 G. P. Smith, Virology 167, 156 (1988).
" S. E. Cwirla, E. A. Peters, R. W. Barrett, and W. J. Dower, Proc. Natl. Acad. Sei. U.S.A. 87, 6378 (1990). d j. j. Devlin, L. C. Panganiban, and P. E. Devlin, Science 249, 404 (1990). e S. Bass, R. Greene, and J. A. Wells, Proteins: Struct., Funct. Genet. 8, 309 (1990). f F. Breitling, S. Dubel, T. Seehaus, I. Klewinghaus, and M, Little, Gene 104, 147 (1991). g A. S. Kang, C. F. Barbas, K. D. Janda, S. J. Benkovic, and R. A. Lerner, Proc. Natl. Acad. Sci. U.S.A. 88, 4363 (1991). h j. Greenwood, A. E. Willis, and R. N. Perham, J. Mol. Biol. 220, 821 (1991).
232
MUTAGENESIS AND GENE DISRUPTION
[15]
and might be strongly selected against in high copy number vectors because of cell killing. A recombinant coat protein gene can be transplanted from the phage genome into a type of plasmid called a p h a g e m i d . 9 Phagemids contain the filamentous phage intergenic region comprising the origins of plus- and minus-strand synthesis and all other cis-acting elements needed for synthesis of ssDNA and packaging it into virions. They also contain a nonphage origin of replication and an antibiotic resistance gene so they can be propagated like any plasmid independently of phage function. When a cell harboring a phagemid is infected by filamentous helper phage, it secretes the phagemid (as well as the helper phage) in the form of infectious virions. These virions display a mixture of recombinant coat protein molecules encoded by the phagemid and wild-type molecules encoded by the helper. Phagemid virions are readily distinguishable from helper virions in that they transduce antibiotic resistance into any cell they infect. In the phagemid pSEX of Breitling et al. ~othe foreign domain is fused to the N terminus of wild-type pill, while in phagemid hGH-MI3glII of Bass et al. 11 the foreign domain r e p l a c e s the exposed N-terminal domain of the wild-type protein. Either design might allow display of foreign domains that would otherwise interfere with infectivity, which does not seem to require that all pill molecules on a particle be functional. Once infected into cells, clones are propagated as ampicillin resistance plasmids independently of pill or pVIII function. Kang et al. 12 (phagemid pCBAK8) and Greenwood et al. t3 (phage fdH and plasmid pKfdH) have spliced foreign inserts into the major coat protein pVIII, resulting in virions with many more displayed copies of the foreign peptide than in the case of gene I I I fusions. Peptides of up to six amino acids can be directly spliced into phage gene VIII, thereby being displayed on all copies of pVIII (2700 in wild-type particles). ~3 By transferring the recombinant gene V I I I to a phagemid or other plasmid and complementing with a helper phage to supply wild-type pVIII, much larger peptides-indeed, a 50-kDa antibody Fab domain12--can be displayed in dozens to hundreds of copies along the length of the virion. Such virions can be used directly as highly effective immunogens for eliciting antibodies against the 9 D. A. Mead and B. Kemper, in "Vectors: A Survey of Molecular Cloning Vectors and Their U s e s " (R. L. Rodriquez and D. T. Denhardt, eds.), p. 85. Butterworths, Boston, 1988. 10 F. Breitling, S. Dubel, T. Seehaus, I. Klewinghaus, and M. Little, Gene 104, 147 (1991). li S. Bass, R. Greene, and J. A. Wells, Proteins: Struct., Funct. Genet. 8, 309 (1990). 12 A. S. Kang, C. F. Barbas, K. D. Janda, S. J. Benkovic, and R. A. Lerner, Proc. Natl. Acad. Sci. U . S . A . 88, 4363 (1991). 13 j. Greenwood, A. E. Willis, and R. N. Perham, J. Mol. Biol. 220, 821 (1991).
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
233
foreign peptides.13 Such a vector could be of particular utility in vaccine development. To create a library, RF or plasmid DNA is restricted and spliced to the foreign DNA insert. Several vectors have two restriction sites yielding single-stranded termini that are neither complementary nor self-complementary; after the short "stuffer" between the sites has been removed, the linear vector DNA cannot be self-ligated, so that circularization requires that a compatible foreign insert be spliced in. TM Table I shows (for the general cloning vectors) the N-terminal primary sequence of the recombinant pill or pVIII encoded by clones carrying a frame-preserving or frame-restoring insert, assuming signal peptidase cleaves in the usual position. The variable part of the sequence, which the user is completely free to specify in the insert, is abbreviated Xn. As can be seen, two of the vectors (fAFF1 and M13LP67) allow the user to specify even the N-terminal residue of the mature gene Ili protein; most amino acids seem to be accepted at the first position after the signal peptide without impairing phage yield or infectivity, 5 although it has not yet been shown that signal peptidase actually cleaves such fusion proteins at the expected position. In the other vectors, insert-specified residues are preceded in the mature protein by a few vector-specified residues. Unless specifically noted, the procedures in this chapter are those used with the fUSE vectors in our laboratory. Although for the most part these procedures exemplify methods in general use with filamentous fusion phage, a few will have to be changed--usually in obvious ways--when applied to other vector systems. Bacterial Strains K-91 is a h- derivative of K-3815; it is Hfr Cavalli and has chromosomal genotype thi. K-91Kan is K-91 with the "mini-kan hopper" element, j6 a kanamycin-resistance transposon without its own transposase gene, inserted into the lacZ gene. K-80217is F - and has chromosomal genotype galK2 galT22 metB1 (lac-3 or lacY1) supE44 hsdr2. MC1061, ]8 the host for electroporation, is F - and has chromosomal genotype hsdR mcrB araD139 A(araABC-leu)7679 AlacI74 galU galK strA thi. z4 A. Aruffo and B. Seed, Proc. Natl. Acad. Sci. U.S.A. 84, 8573 (1987). 15 L. B. L y o n s and N. D. Zinder, Virology 49, 45 (1972). 16 j. C. Way, M. A. Davis, D. Morisato, D. E. Roberts, and N. Kleckner, Gene 32, 369 (1984). 17 W. B. Wood, J. Mol. Biol. 16, 118 (1966). ]8 p. S. Meissner, W. P. Sisk, and M. L. B e r m a n , Proc. Natl. Acad. Sci. U.S.A. 84, 4171 (1987).
234
MUTAGENESIS AND GENE DISRUPTION
[15]
Recipes ABTS solution: Dissolve 2,2'-azinobis(3-ethylbcnthiazoline-6-sulfonic acid) (Sigma, St. Louis, MO) at 0.22 mg/ml in a 386 : 614 (v/v) mixture of 0.2 M Na2HPO 4 and 0.1 M citric acid; then add 1/1000 vol of 30% (w/v) H202 and use within 15 min Avidin-peroxidase complex: ABC reagent (Vector Laboratories, Burlingame, CA) is prepared in TBS (see below) with 0.1% (v/v) Tween 20 as recommended by supplier. Presumably other avidin-peroxidase or streptavidin-peroxidase complexes can be substituted Bovine serum albumin (BSA): Unless otherwise indicated, this is bovine serum albumin fraction V (Sigma); the protein is dissolved at 50 mg/ml in water, filter-sterilized, and stored at 4 or - 2 0 ° Acetylated BSA: This nuclease-free protein, which is used as a carrier for DNA reactions, is purchased from Promega (Madison, WI) as a l-mg/ml stock and stored at - 2 0 ° Dialyzed BSA: We use dialyzed fraction V BSA (Sigma) as a carrier protein when small amounts of contaminating biotin might interfere; a 50-mg/ml stock solution is prepared and stored as for nondialyzed BSA Blocking solution: 5 mg/ml dialyzed BSA, 0.1 M NaHCO3, 0.1 tzg/ml streptavidin, 0.02% (w/v) NaN3 ; filter sterilized and stored in refrigerator; can be reused many times Blotto solution: 5 g nonfat dry milk in 100 ml TBS/azide (see below) Bonding coat: Mix 20 tzl 7-methacryloxypropyltrimethoxysilane (Sigma) with 20 m195% (v/v) ethanol, then add a mixture of 60 tzl glacial acetic acid and 600 tzl water; use within 1 hr Buffered glucose: 50 mM glucose, 25 mM Tris-HCl (pH 8.0), 10 mM EDTA, pH 8 (see below). Store in refrigerator EDTA, pH 8 (250 mM stock): 0.25 M ethylenediaminetetraacetic acid disodium salt (N%EDTA), pH adjusted to 8.0 with NaOH; the stock is autoclaved and stored at room temperature Elution buffer: 0.1 N HC1 adjusted to pH 2.2 with glycine, 1 mg/ml BSA; store in refrigerator. The glycine hydrochloride buffer is made as a 4 × stock, filter sterilized, and stored at room temperature Ethanol: 100% ethanol is poured into a sterile glass-stoppered bottle and used directly; 70% (v/v) ethanol is 64.9% (w/w) ethanol in sterile water Formamide load buffer: Mix 6.65 ml formamide, 280 IA 0.5 M Na2EDTA (pH adjusted to 8.0 with NaOH), 3.5 mg bromphenol blue, 3.5 mg xylene cyanol FF, and 70 tzl water; store at - 2 0 ° GB B (40 × stock): 1.68 M Tris, 0.8 M sodium acetate, 72 mM Na2EDTA, pH adjusted to 8.3 with glacial acetic acid; store at room temperature
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
235
Kinase buffer (10 × stock): 0.5 M Tris-HCl (pH 7.5), 100 mM MgC12, 50 mM dithioerythritol, 1 mM spermidine, 1 mM EDTA (pH 8). Store at - 2 0 °, thaw, and refreeze as needed LB medium: See Sambrook e t a l . 19 Ligation buffer (5 × stock): 150 mM Tris-HCl (pH 7.5), 10 mM dithiothreitol (DTT), 1 mM EDTA (pH 8), 5 mM spermidine, 1.25 mM ATP (added as 100 mM stock neutralized to pH -7), 150 mM NaCI, 37 mM MgCI 2 , 500-800/xg/ml acetylated BSA NaN 3 (5% stock): A 5% (w/w) solution is made up, taking precautions to avoid exposure to the toxic chemical; store in refrigerator NAP buffer: Autoclave 90 ml 88 mM NaC1; when cool add 10 ml sterile 0.5 M NH4H2PO 4 (pH adjusted to 7.0 with NH4OH; made as 0.5 M stock, which is separately autoclaved in tightly closed bottle and stored at room temperature) NZY medium: Same as NZYM (Sambrook e t al. 19) without added MgSO4; autoclave and store at room temperature NZY agar medium: For 1 liter (about forty 100-mm petri dishes) autoclave 11 g Bacto-agar (Difco, Detroit, MI) in 500 ml water in a 2-liter plastic flask; while still hot add 500 mi sterile 2 x NZY at room temperature (and antibiotics and other heat-labile components as appropriate), mix by gentle swirling, and pour - 2 5 ml per 100-mm petri dish NZY/Tc agar medium: NZY/agar medium with Tc at 40/~g/ml NZY/Tc medium: NZY with 20/zg/ml Tc PEG/NaC1 solution: Mix 100 g polyethylene glycol (PEG) 8000 (Fisher, St. Louis, MO), 116.9 g NaCI and 475 ml water, heating if necessary to dissolve all the solid; store at room temperature or in refrigerator. Can be autoclaved, but must be mixed occasionally as it cools to prevent separation of phases Sequencing gel solution: The 40% (w/w) acrylamide stock is 38% (w/w) acrylamide, 2% (w/w) bisacrylamide; it is made up taking precautions to avoid exposure to the neurotoxic monomer, and stored up to - I year in the refrigerator. The 6% acrylamide gel solution (500 ml) is made by mixing 75 ml 40% acrylamide stock, 125 ml water 250 g ultrapure urea, and 100 ml 5 x TBE buffer (see below); the solution is filtered through a 0.45-~m nitrocellulose filter, degassed, and stored for up to - 2 months at room temperature in an amber bottle SOC medium: See Sambrook e t a l . 19
19 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2nd Ed., Vols. 1-3. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989.
236
MUTAGENESIS AND GENE DISRUPTION
[15]
Sodium acetate buffer (3 M stock): 3 M sodium acetate adjusted to pH 6.0 with acetic acid; autoclave with cap tight, store at room temperature Soft agar: Autoclave 1 g Bacto-tryptone, 0.5 g NaC1, and 0.75 g Bactoagar in 100 ml water; store at room temperature; for use, melt in microwave oven and dispense -3-ml portions into sterile 13 × 100 mm tubes in a 50° heating block as needed TBE buffer (5 x stock): 0.5 M Tris, 0.5 M H3BO3, I0 mM NazEDTA; store at room temperature TBS (10 x stock): 1.5 M NaCI, 0.5 M Tris-HCl (pH 7.5); dilute and autoclave as needed; both the 1 × and 10x solutions are stored at room temperature TBS/Tween: make by diluting 0.5 ml Tween 20 in 100 ml TBS; autoclave and store at room temperature TBS/gelatin: 0.1% (w/v) gelatin in TBS; autoclave and store at room temperature TBS/azide: TBS with 0.02% (w/v) NaN3 ; store at room temperature TE: l0 mM Tris-HCl (pH 8.0), 1 mM EDTA (pH 8); make as 10 × stock, dilute and autoclave as needed; store at room temperature Termination mixes and termination diluent: Diluent contains 67 ~M dNTPs (purchased from Pharmacia, Piscataway, N J, as 100 mM stock solutions neutralized to pH -7), 16.7 mMTris-HC1 (pH 7.5), 66.7 mM NaCI, 13.3 mM dithioerythritol, and 100 ~g/ml acetylated BSA. The Q, R, W, M, K, and S mixes contain in addition the following ddNTPs (purchased from Pharmacia as 5 mM stock solutions neutralized to pH -7): R, 3.2/~M ddATP and ddGTP, 0.32/~M ddCTP; Q, 3.2/~M ddTTP and ddATP, 0.32/~M ddCTP; W, 3.2/~M ddATP and ddTTP; M, 3.2/~M ddATP and ddCTP; K, 3.2/~M ddGTP and ddTTP; S, 3.2 /~M ddGTP and ddCTP. All these solutions are stored at - 2 0 ° and thawed and refrozen as needed Tetracycline (Tc) (20-mg/ml stock): Dissolve solid in water at 40 mg/ml, filter sterilize into an equal volume of autoclaved, cooled glycerol, mix, and store at - 2 0 ° General Procedures
Phenol Extraction, Chloroform Extraction, and Ethanol Precipitation from Sodium Acetate Solution The steps are carried out as outlined by Sambrook et al. 19 Removal of Supernatant. In some procedures it is advantageous to remove almost all supernatant from a pellet; to accomplish this, decant or aspirate the supernatant, recentrifuge the tube (maintaining the same
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
237
orientation in the rotor), and aspirate or pipette off the residual supernatant that is thus driven to the bottom of the tube. Electroporation. Large-scale transfections for library construction are accomplished by electroporation. 2° Frozen Escherichia coli MCl061 cells 2° (40-50 /~l) are thawed, mixed with DNA (up to - l p~g in 1-5 p,l water), transferred to a cold 2-ram cuvette (Bio-Rad Laboratories, Richmond, CA) connected in parallel with a 400-f~ resistor, and shocked by charging a 25-/~F capacitor to 2.5 kV and discharging it through the resistor. The shocked cells are immediately suspended in SOC medium containing 0.2/~g/ml Tc, transferred to a sterile 15-ml culture tube, and shaken at 37 ° for 30-60 min before spreading 200-/~1 portions on NZY/Tc agar medium in 100-mm petri dishes or inoculating into liquid NZY/Tc medium. Small-scale transfections (e.g., for strain constructions) are accomplished by the CaCl 2 method as described. 21 Vector DNA. Noninfective phage (e.g., fUSE vectors) should be propagated in an F - (uninfectable) host to guard against accumulation of infective pseudorevertants; infective phage can be grown in male (F +, F', or Hfr) strains. Cultures are grown to stationary phase in 1 liter of LB medium containing 15 /.Lg/ml Tc. Cells are suspended in 40 ml buffered glucose and RF is extracted by alkaline lysis (Sambrook et al.19), except that no lysozyme is required. The crude DNA, dissolved in l0 ml TE, is extracted twice with phenol and once with chloroform, reprecipitated with ethanol, and purified by CsCl-ethidium bromide density gradient centrifugation (Sambrook et al., 19pp. 1.42-1.43); one tube for the VTi50 rotor (Beckman Instruments, Fullerton, CA) accommodates DNA from up to 2 liters of culture; because of the large amount of RNA in the lysate, at least 25 mg ethidium bromide should be included to ensure enough free dye to saturate the DNA. Purified DNA is cleaved with appropriate restriction enzyme(s) and extracted with phenol and chloroform. To remove a "stuffer" fragment released by cleavage at two restriction sites, the cleaved DNA (37/~g/ml in restriction buffer or a dilution thereof) is precipitated by adding 1/9 vol 3 M sodium acetate buffer and 2/3 vol 2-propanol, incubating on ice 20 min, microfuging 30 rain at room temperature, removing all supernatant (see Removal of Supernatant, above), washing gently in 70% (v/v) ethanol, and again removing all supernatant. The precipitate is dried briefly under vacuum, dissolved in TE, ethanol precipitated, and quantified spectrophotometrically. Large-Scale Purification o f Virions. The following procedure is for derivatives of fd-tet, which yield about 5 x l0 ll particles/ml of culture. 20 W. J. Dower, J. F. Miller, and C. W. Ragsdale, Nucleic Acids Res. 16, 6127 (1988). 2i F. K. Nelson, S. M. F r i e d m a n , and G. P. Smith, Virology 108, 338 (1981).
238
MUTAGENESIS AND GENE DISRUPTION
[15]
One liter of stationary-phase culture (previous section) is freed of cells by two successive 10-min centrifugations [5000 and 8000 rpm in three 500-ml bottles in the Sorvall (Norwalk, CT) GS3 rotor at 4 °]; the final supernatant is distributed equally in three 500-ml centrifuge bottles. (If needed, RF can be prepared from the cell pellet, as described in the previous section.) After adding 0.15 vol PEG/NaC1 solution, the supernatnat is thoroughly mixed and incubated overnight in the refrigerator. Precipitated virions are collected by a 30-min centrifugation (8000 rpm in the Sorvall GS3 rotor at 4°), removing all supernatant. The precipitate is dissolved and pooled in a total of 30 ml TBS, transferred to a single Oak Ridge tube (Nalge, Rochester, NY), cleared by a 10-min, 15,000-rpm centrifugation (Sorvall SS34 rotor), and reprecipitated from the supernatant by adding 4.5 ml PEG/ NaCI solution and incubating at least 1 hr in the refrigerator. Virions are collected by centrifugation (again taking care to remove all supernatant), dissolved in 10 ml TBS, and cleared by centrifuging at 15,000 rpm for 10 rain at 4 ° (Sorvall SS34 rotor). The final supernatant is transferred to a tared vessel and TBS added to bring the total net weight to I0.75 g; 4.83 g CsC1 is added to bring the density to 1.30 g/ml. The CsC1 solution is transferred into a ~ × 2½ in. polyaUomer tube (topping with mineral oil if necessary) and centrifuged at 37,000 rpm for at least 40 hr in the SW41 rotor (Beckman Instruments) at 4 °. (Note: For phage other than fd-tet derivatives, the particle yield is - 2 × 1012virions/ml, and the virions from 1 liter of culture should be distributed into three or four SW41 tubes.) After centrifugation light-scattering bands are visualized by shining a bright light downward through the tube and looking through the wall of the tube at right angles to the light beam. The translucent, nonflocculent phage band lies near the middle of the tube just above a sharp, white, flocculent band; 1015 particles can give a band - I cm wide. The phages are collected by puncturing the side of the tube with a 16-gauge needle attached to a 3-ml syringe. Phages from one or two tubes are transferred to 26-ml polycarbonate bottles for the Beckman 60Ti rotor; the bottles are filled with TBS and centrifuged at 50,000 rpm for 4 hr at 4° to pellet virions. After removing all supernatant (see Removal of Supernatant, above), the virion pellet in each bottle is dissolved in 3.2 ml TBS, divided into three 1.5-ml microcentrifuge tubes (800/zl each), and cleared by microfuging 3 min; supernatants are transferred to polyallomer microcentrifuge tubes for the Beckman TLA-100.1 rotor and centrifuged at 57,000 rpm for 90 mip. at 4 °. After again removing all supernatant, the pellets from 1 liter of original culture are dissolved and pooled in a total volume of 3 ml TBS. Virions are quantified spectrophotometrically by scanning a 1/50 dilution from 240 to 320 nm, giving a broad peak at 260-280 nm with peak absorp22 L. A. Day, J. Mol. Biol. 39, 265 (1969).
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
239
tion at 269 n m . 22 Virion concentration (physical particles per milliliter) can be estimated a s A269 X 6 X 10t6/(bases/ssDNA); a typical yield for fd-tet derivatives is 3 x 10 u virions/ml of culture supernatant. Starved Cells. Filamentous phages infect E. coli strains displaying the sex pilus encoded by the F episome (Hfr, F*, or F' strains). Because phages adsorb slowly to cells at log-phase concentrations ( - 5 x 10s cells/ ml), we concentrate cells to - 10~°cells/ml after starvation. 23 Cells are first grown in 20 ml NZY medium at 37 ° with vigorous shaking to an optical density of - 0 . 4 - 0 . 6 at 600 nm; shaking is then slowed for 5 min, and care is taken in subsequent steps to avoid shearing the fragile F pili. The culture is centrifuged (2200 rpm for 10 rain in a Sorvall SS34 rotor), the supernatant is poured off, and the cells are gently resuspended in 20 ml 80 mM NaCI and shaken gently in a 125-ml culture flask at 37 ° for 45 rain to starve the cells. The cells are collected as above and resuspended in 1 ml cold NAP buffer; they can be stored in the refrigerator for about 5 days without affecting titers. Titering Transducing Units. Infections are ordinarily carried out either in disposable 15-ml culture tubes (if only a few infections are involved) or in wells of a 24-well culture dish (Costar, Cambridge, MA). To 20/zl phage (appropriately diluted in TBS/gelatin) is added 20/zl starved cells; after 10 min at room temperature the mixture is diluted with 0.2-2 ml NZY medium containing 0.2/xg/ml Tc and incubated with shaking (gentle shaking in the case of 24-well culture dishes) for 20-40 min. These cultures or appropriate dilutions of them (using NZY with or without 0.2 gg/ml Tc as diluent) are then spotted (20/.d/spot) or spread (50-200/zl/100-mm dish) on NZY/Tc agar medium. Colonies are counted after 16-36 hr at 37 °, each colony representing a transducing unit (TU). Biotinylation. The following protocol is used for biotinylating antibodies and antibody Fab fragments; it may have to be modified for other ligates. The protein must be freed of primary and secondary amines, azide. and strong buffers that would prevent pH adjustment in the next step. Antibody (10-50/~g in 20/xl) is brought to pH 8-9 in a siliconized 1.5-ml microtube by adding 4.4/zl 1 M NaHCO3. Sulfosuccinimidyl-6-(biotinamido)hexanoate (NHS-LC-biotin, M r 556.6; Pierce Chemical, Rockfold, IL) is dissolved at 0.5 mg/ml (0.88 mM) in 2 mM sodium acetate buffer and 20/xl is immediately added to the antibody; the NHS-LC-biotin, which is freely water soluble because of the sulfo group, is slowly depleted by hydrolysis in aqueous solution. Coupling is allowed to progress 2 hr at room temperature, then quenched by adding 200/~1 1 M ethanolamine (pH adjusted to 9.0 with HCI) and incubating two additional hours at room temperature. Carrier protein (20/xl 50-mg/ml dialyzed BSA) is added and 2~ H. Tzagoloff and D. Pratt,
Virology 24, 372 (1964).
240
MUTAGENESIS AND GENE DISRUPTION
[].5]
the reaction mixture is diluted with 1 ml TBS and concentrated and washed twice with TBS and once with TBS/azide on a 30-kDa Centricon ultrafilter (Amicon, Danvers, MA) according to the directions of the supplier. The concentration of biotinylated antibody is calculated from the volume of the final concentrate (usually 35-50/~1), assuming no loss. Processing Small Cultures. Colonies are inoculated into 1.5 ml NZY/ Tc medium in 15-ml disposable culture tubes or in 18 x 150 mm glass culture tubes. The tubes are secured vertically in a shaker incubator and shaken vigorously at 37° for 12-24 hr. Each culture is poured into a 1.5-ml microtube and microfuged - 1 min to pellet cells; 1 ml of the culture supernatant is pipetted into a fresh microtube containing 150/xl PEG/NaC1 solution, and after thorough mixing virions are allowed to precipitate overnight in the refrigerator. Precipitated virions are collected by microfuging (15 min at - 4 ° or at room temperature) and removing all supernatant (see Removal of Supernatant, above), and dissolved in 1 ml TBS; virion concentration is - 5 x 10 u particles/ml. A sequencing template can be prepared from 200/.d of this solution. Alternatively, if phage are to be used for enzyme-linked immunosorbent assay (ELISA) they can be further purified as described in the next section. Partial Purification of Phage for ELISA. The dissolved first PEG precipitate (1 ml; see previous section) is cleared by microfuging - 1 min, and supernatant pipetted into a fresh microtube containing 150 txl PEG/NaCI solution. Phage are precipitated and collected as in the previous section, and dissolved in 110/xl unbuffered 0.15 M NaCI. The solution is cleared by microfuging - 1 min, and 100 ~1 of the supernatant is transferred to a fresh microtube containing 11 /xl 1 M acetic acid; after I0 min at room temperature and 10 min on ice, acid-precipitated virions are collected by microfuging 30 min at - 4 ° and removing all supernatant. The final virion pellet is dissolved in 500 txl TBS, giving a phage concentration of - 5 x 10u particles/ml. A sequencing template can be prepared from 200/xl of this solution. Sequencing Template. Phages (200 tzl containing -1011 particles purified by at least one PEG precipitation, as in previous two sections) are extracted once with phenol and once with chloroform in 500-tzl microtubes. The final aqueous phase (100-150 /xl) is transferred to a 1.5-ml microtube containing 250 ~1 TE and 40 ~1 3 M sodium acetate buffer (see Recipes, above), and the viral ssDNA precipitated by adding 1 ml ethanol. After at least 1 hr at 4 °, ssDNA is collected by microfuging 30 min, aspirating the supernatant, gently adding - 1 ml 70% (v/v) ethanol, and removing all supernatant; the DNA is dried briefly in vacuo, dissolved in 7.5/xl water, and stored at - 2 0 ° until use. Sequencing Reactions. We use an 18-base primer complementary to the wild-type gene III sequence 18-32 bases 3' of the cloning sites in the
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
241
fUSE vectors. The primer is end labeled by mixing 267 ng of the 5'-OH oligonucleofide with 200/xCi [y-32p]ATP [specific activity -7000 Ci/mmol; crude preparation from ICN Biochemicals (Costa Mesa, CA) is used within - 4 weeks of reference date], 1 /zl 10 x kinase buffer, and 1 /zl T4 polynucleotide kinase, in a total volume of 10/zl in a 1.5-ml microtube; incubating at 37° for 30 min; and adding 125/zl TE and heating to 65-70 ° for 15 min to inactivate the enzyme. If sequences are to be "piggy-backed," this end-labeled primer is freed of unincorporated isotope and other impurities by adsorption to NENSORB 20 (New England Nuclear, Boston, MA), washing with adsorption buffer and water, and eluting with a 1 : 1 (v/v) ethanol-water mixture as in the supplier instructions; the radioactive fractions are dried in vacuo and dissolved in water. Water and stock solutions are added to the labeled primer (whether NENBSORB purified or not) to give final concentrations of 128 mM Tris-HC1 (pH 7.5), 160 mM NaCI, and 48 mM DL-isocitrate (pH adjusted to 7-7.5 with NaOH) in a final volume of 200-900/zl; this solution is called 2 x primer-buffer, and is stored for up to 4 weeks at - 2 0 °. Just before use, the required volume of 2 x primer-buffer is mixed with an equal volume of 16 mM MnCI 2, and 6.7 p~l of the mixture is added to sequencing template in 7.5/~1 water (previous section); after microfuging briefly to mix the solutions, the mixture is heated at 65-70 ° for 5-10 min, then allowed to cool gradually to room temperature over a period of at least 30 min; the resulting primed templates can be stored in the refrigerator or freezer for at least 4 days, during which time they turn brown without apparent ill effects. The primed template is microfuged briefly to drive condensation to the bottom, and 3-/zl droplets are deposited in a grid pattern on a polystyrene petri dish--two droplets (R and Q) per template for two-lane sequencing, four droplets (W, M, K, and S) for four-lane sequencing. Up to 40 droplets fit easily in a single 100-mm dish; it is convenient to deposit droplets to be loaded on the sequencing gel at different times on different dishes. T7 DNA polymerase (e.g., Sequenase version 1 or 2 from United States Biochemical, Cleveland, OH) is added to the appropriate volume of one of the termination mixes (diluted with termination diluent if appropriate) in a siliconized microtube to give a concentration of 0.25 units//zl; 3-/zl portions of the mixture are immediately deposited on each of the appropriate droplets on the petri dish. This process is repeated with the remaining termination mix(es), and the petri dish is floated on a 37 ° water bath for 5-10 min. Polymerization reactions are stopped by depositing a 4-/.d droplet of formamide load buffer on each droplet. The petri dish can be stored at - 2 0 ° for several days if convenient. Pouring and Running Sequencing Gels. Sequencing gels (38 cm long, 30.5 cm wide, and 0.4 mm thick) are run in a standard sequencing gel
242
MUTAGENESIS AND GENE DISRUPTION
[15]
apparatus (e.g., model $2; Bethesda Research Laboratories, Gaithersburg, MD). The short glass plate is coated twice with - 7 ml Rain-X 24 (Unelko, Scottsdale, AZ); after each coat has been applied with a Kimwipe, the reagent is allowed to dry, then the plate is wiped with 95% (v/v) ethanol and buffed with a Kimwipe. Meanwhile, the long plate is coated three times with - 6 ml bonding coat (see Recipes, above); after each coat has been applied with a Kimwipe, the reagent is allowed to dry and the plate is buffed with a Kimwipe. It is crucial not to cross-contaminate the plates. The acrylamide bonds covalently to the coated long plate, and after electrophoresis can be dried onto the glass surface without cracking. Plates are assembled into a sandwich with a spacer along the bottom as well as each side, and clamped together with clips; there is no need to seal the sandwich with tape. Sequencing gel solution (100 ml) is measured into a beaker and mixed with 1 ml 10% (w/w) freshly dissolved ammonium persulfate. A 3-ml portion is mixed with 3/xl N,N,N',N'-tetramethylethylenediamine (TEMED) in a test tube and immediately pipetted down one edge of the sandwich, which is laid flat as the gel polymerizes ( - 5 min), sealing the edges and bottom of the sandwich. Then 21 /zl TEMED is added to the remainder of the gel solution, which is mixed, poured into a 250-ml squirt bottle, and injected between the glass plates. The nonserrated eges of sharkstooth combs are inserted to create a flat surface, and the gel is polymerized 2-24 hr. The gel is assembled into the electrophoresis apparatus with sharkstooth combs, creating either forty-nine 6-mm or ninety-seven 3-mm wells, and preelectrophoresed until the surface temperature is - 5 0 °. Just before loading, the petri dish containing the reactions (previous section) is floated with its lid off on a water bath at 80-90 ° for 3 min, thus denaturing the DNA and evaporating the samples to - 4 - 5 / x l . After wells are rinsed to remove urea, samples are loaded (4 /zl/6-mm well, 2/zl/3-mm well) and the gel is electrophoresed until the xylene cyanol FF (blue-green) band has migrated 21 cm for a normal 6% (w/v) gel. On a piggy-back gel, xylene cyanol FF is run to 14 cm, then a second set of samples is loaded and run until the xylene cyanol FF from the first stample has run 24 cm. This allows reliable reading of strands from the first load with lengths between 55 and 80 bases (including the primer), without interference from primer breakdown products from the second load. This is adequate for clones from our hexapeptide epitope library, in which the 18-residue sequence encoding the variable hexapeptide corresponds to strand lengths between 61 and 78 bases. After electrophoresis, the gel, which is bonded to the long glass plate, 24 R. S. Barnett and J. N. Davidson, in " F o c u s , " (N. Sasavage, ed.) Vol. 11, p. 75. Bethesda Research Laboratories, Bethesda, Maryland, 1989.
[15]
LIBRARIES OF P E P T I D E S D I S P L A Y E D BY P H A G E
243
is washed twice with 600-1000 ml 10% (v/v) methanol, 10% (v/v) acetic acid (the second wash can serve as the first wash of the next gel), dried onto the plate (e.g., - 3 hr at room temperature under a current of air), and autoradiographed at room temperature. Overnight exposure is usually sufficient for short sequences with undiluted termination mixes; 24-48 hr usually suffices for longer sequences. After use, the long glass plate is soaked in 1 N NaOH to remove the gel and cleave off the bonding coat; the gel is scooped into radioactive waste, while the alkaline solution can be reused several times. The plate is then rinsed and recoated as described above. Making a Library
Design of Degenerate Oligonucleotide Insert Most libraries to date have used synthetic degenerate oligonucleotides as the insert. A degenerate oligonucleotide is a mixture of sequences created in a single synthesis by coupling mixtures of nucleotides, rather than single nucleotides, at selected positions in the growing chain. At the codon level degeneracy can be of two general types: fully degenerate codons encode all 20 amino acids with no bias beyond what is entailed by the unequal degeneracy of the genetic code; while doped codons are biased toward one particular amino acid in order to introduce random substitutions into a base peptide or polypeptide sequence. The first two positions of each fully degenerate codon are synthesized by adding an equimolar mixture of dA, dC, dG, and T to the growing oligonucleotide; the third position has an equal mixture of dG and T (dG and dC are also acceptable). The resulting mixture of 32 triplets encodes all 20 amino acids, and includes only the amber chain termination codon. Doped codons are synthesized by doping each nucleotide encoding the base peptide sequence with a mixture of the other three nucleotides. For instance, a nucleotide substitution rate of 30% results if, during synthesis, 60% of the nucleotide in each step is specified and 40% comes from an equimolar mixture of all four nucleotides. Whatever its design, the inserted coding sequence must fuse correctly to the coat protein reading frame at both ends of the cloning site in order to retain or restore gene function and thereby be expressed on the virion.
Construction of Library Efficient cloning of oligonucleotides with long degenerate sequences presents a special problem. Even if as few as six degenerate codons are
244
MUTAGENESIS AND GENE DISRUPTION
[15]
represented, the overall complexity exceeds that of the entire human genome, and it is not practical to create a double-stranded degenerate insert simply by annealing complementary degenerate oligonucleotides. Two ways of circumventing degeneracy have been used: polymerase chain reaction (PCR) amplification of a single-stranded degenerate oligonucleotide template with primers corresponding to nondegenerate flanking sequences3; and creation of "gapped duplexes," in which the degenerate region remains as a single-stranded gap in an otherwise double-stranded insert. 5 To synthesize a double-stranded degenerate insert by PCR, nondegenerate sequences with appropriate restriction sites are positioned on both sides of the degenerate codons in the chemically synthesized template. The template is PCR amplified using 5'-biotinylated primers corresponding to the flanking regions. A 1-ml PCR mixture containing -40 pmol template and 750 pmol of each biotinylated primer is incubated with 25 units of AmpliTaq DNA polymerase (Perkin-Elmer-Cetus, Norwalk, CT) in the buffer recommended by supplier; after five temperature cycles (2.5 min at 95°, 4 min at 42°, 4.4 min at 72°) and 5 min at 72°, polymerization is stopped by the addition of 4/A 250 mM EDTA, pH 8. The product is precipitated with ethanol and dissolved in 100/A TE and crudely quantified by gel electrophoresis next to appropriate double-stranded oligonucleotide standards. A 300-ng portion is digested in a 500-/B reaction mixture containing 1200 units of a restriction enzyme that will produce appropriate overhanging ends for splicing into linearized vector DNA. The digestion is stopped by addition of 22 /A 250 mM EDTA, pH 8, and mixed with streptavidin-agarose beads [Bethesda Research Laboratories; 200/A 50% (w/v) suspension] that have previously been washed l0 times with 0.1 M NaCl in TE by suspension and centrifugation. After 30 min of gentle agitation, the beads are centrifuged and the supernatant is removed to another tube. Beads are washed twice with 200/A water, and the supernatants are pooled with the main supernatant. The final product is extracted with phenol and chloroform, concentrated by evaporation in vacuo, and quantified roughly by gel electrophoresis. The ligation reaction contains 5 p,g/ml linearized vector DNA, a twofold molar excess of double-stranded degenerate insert, and 10 units/ml T4 DNA ligase in ligation buffer. After incubation at 20° for 12-18 hr, products are analyzed by running a 20-/A portion of the ligation mixture on a 0.8% (w/v) agarose gel containing 4 x GBB and 1.5 /xg/ml ethidium bromide. The product usually consists of a mixture of nicked circular, covalently closed circular, and residual linear RF DNA, which comigrate with linear double-stranded marker fragments of -11, - 5 , and 9.2 kbp, respectively. A good ligation produces - 2 × l07 Tcresistant transfectants per microgram vector input by electroporation.
[15]
L I B R A R I E S O F P E P T I D E S D I S P L A Y E D BY P H A G E
245
In the gapped duplex method, 5 a long degenerate oligonucleotide is annealed to two short oligonucleotides ("patches") complementary to nondegenerate regions at its 5' and 3' ends. When ligated to linearized vector DNA (essentially as described above), this gapped duplex produces transfectants at about the same frequency as do double-stranded inserts. When the degenerate region is short (e.g., six codons), most phage in a library produced by either procedure have inserts with the correct reading frame and sequence. When long degenerate sequences (e.g., 20 codons) are cloned as gapped duplexes, however, the incidence of rearranged sequences increases (J. K. Scott, unpublished observations, 1991); this problem can be at least partially ameliorated by filling in the long singlestranded gaps with DNA polymerase plus ligase. The ligation product is transfected into cells by 50-100 separate electroporations; the transfected cells are propagated in 10 to twenty l-liter flasks containing 100 ml NZY/Tc medium (5 electroporations/flask) at 37 ° with vigorous shaking for 24 hr. The cultures are pooled and the virions purified and quantified spectrophotometrically (see Large-Scale Purification of Virions, above); the concentration of infectious particles is determined by titering transducing units. Gross clonal bias and sequence abnormalities can be detected by sequencing individual clones of electroporated cells. Affinity Purification of Phage from Epitope Library
Biopanning This affinity purification technique z relies on the superstrong biotin-streptavidin reaction to attach ligate-binding phage to a solid surface (Fig. 1). Alternative techniques are discussed in the next section. Into a 60-mm polystyrene petri dish is pipetted 2 ml 0.1 M NaHCO3 ; 20 tzl 1-mg/ ml streptavidin is pipetted into the buffer, and the dish is jiggled until the entire bottom surface remains wetted. The dish is covered with the lid and kept overnight at - 4 ° in a humid chamber, preferably on a rocker. The next day the lid is discarded, the streptavidin solution is poured off, and the dish is filled with blocking solution (see Recipes, above) and rotated or rocked 1-2 hr at room temperature. The blocking solution is poured back into its container for reuse, and the dish is washed three times by filling it with TBS/Tween, pouring out the buffer, and slapping the dish on a clean paper towel (the last wash is not poured off until just before the ligate-reacted phages are added). Meanwhile, a phage mixture--typically 5 ~1 of a library--is reacted overnight with up to 1/xg biotinylated ligate (typically in 5/zl TBS/Tween) in a 1.5-ml microtube in the cold. The reaction is diluted with 1 ml TBS/
246
MUTAGENESIS AND GENE DISRUPTION
[15]
() Oligo Insert encoding epltope
q
Epltope ( ~.
Blotlnylated
antibody
Streptavldin-coated petri dish FIG. 1. Biopanning. Phages in the epitope library are represented with different oligonucleotide inserts in their DNA and the corresponding peptides displayed on pill at one tip of the virion; only one recombinant pill molecule is shown, although there are actually five per particle. Phages displaying a ligand for the biotinylated ligate (an antibody in this example) are captured on the dish by binding of the biotin moiety to immobilized streptavidin.
Tween and immediately pipetted onto the streptavidin-coated dish, which is rocked at room temperature for 10-15 min; during this incubation, phages whose displayed ligands have bound to the biotinylated ligate are in turn attached to the plastic surface via biotin-streptavidin bonds (Fig. 1). The fluid is poured out, and the plate is washed 10 times with TBS/Tween, each time rocking the dish a few minutes, decanting the wash, and slapping the dish face down on a clean paper towel to remove residual wash; the entire process takes 15-60 min. Bound phages are eluted by pipetting 800/zl of elution buffer (see Recipes, above) into the dish, rocking it 5-15 min, and pipetting the eluate into a 1.5-ml microtube containing 48/xl 2 M Tris (pH unadjusted), raising the pH of the eluate to - 8 . The first round of biopanning is critical to success. Ordinarily each clone in the original library is represented by only - I 0 0 infectious units (TU) in the initial 5-p.l aliquot. Because biopanning gives only a 1% yield with strongly binding phage, many binding clones will be represented (if at all) by a single TU. That is why a large amount of biotinylated ligate
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
247
(1 g.g) is used in the first round, to maximize yield even at the cost of reduced discrimination. Ordinarily a single round of biopanning is not sufficient to purify binding clones from a large, complex library. Consequently, phages in the first eluate are biopanned again after amplification. Because it is important to represent as many of the phages as possible (previous paragraph), the eluate is first concentrated (and washed once or twice with TBS) on a Centricon 30-kDa ultrafilter so that the entire eluate can be amplified. The retentate ( - 5 0 tzl) is mixed with 50 /xl starved cells (usually K-91 or K-91Kan; no more than I day old; see General Procedures, above) in a sterile 15-ml tube; after 10 min at room temperature, 2 ml NZY medium containing 0.2/xl/ml Tc is added and the tube is incubated at 37 ° (with shaking if possible) for 45-60 min. The culture is then spread on ten 100-mm dishes (200/zl/dish) containing NZY/Tc agar medium and grown 24-48 hr at 37°; usually there are 500-5000 colonies per dish. The phages (along with the cells that have secreted them) are harvested from the petri dish by scraping them into a total of 30 ml TBS. The suspension is cleared by two centrifugations (5000 and 10,000 rpm, 10 min each, in Oak Ridge tubes in the Sorvall SS34 rotor), and the phages are precipitated from the cleared supernatant in an Oak Ridge tube by adding 0.15 vol of PEG/NaCI solution, mixing thoroughly by multiple inversions, and chilling in the refrigerator at least 4 hr. The precipitate is collected by centrifugation (10,000-15,000 rpm for 15 min in the Sorvall SS34 rotor), and after removing all supernatant (see Removal of Supernatant, above) the pellet is dissolved in 1 ml TBS and transferred to a 1.5-ml microtube. The phage solution is cleared by microfuging a few minutes, the supernatant is transferred to a fresh microtube, and the phages are reprecipitated by adding 150 tzl PEG/NaC1 solution, mixing thoroughly by multiple inversions, and chilling in the refrigerator at least 1 hr. The precipitate is collected by microfuging 5-10 min, all supernatant is removed, the pellet is dissolved in 100/xl TBS, and the solution is cleared by microfuging a few minutes and transferring the cleared supernatant to a fresh microtube. This amplified eluate typically has a titer of 10W 1013TU/ml; the yield is not dependent on the number of colonies per dish, as long as that number exceeds -500. A 5-/~1 portion of the amplified eluate is biopanned in the same manner as the original library. If binding clones predominate in the eluate, there can be as much as 4 pmol phage-borne ligand in 5 ~1 of amplified eluate. By using substantially less than 4 pmol ligate in the second round of biopanning, it is possible to set up a competition among binding clones, thus potentially increasing discrimination. This reduces the yield of binding clones, of course; but because 5 ~1 of amplified first eluate will contain at least 105 TU/TU in the original unamplified first eluate, high yields are
248
MUTAGENESIS AND GENE DISRUPTION
[15]
not nearly as important as in the first round of biopanning. Usually a 50/zl portion of the second eluate (unconcentrated) is amplified as described above, except that only 1 ml of NZY medium with 0.2/zg/ml Tc is added, the infected cells are grown in only five 100-mm petri dishes, and the final PEG pellet is dissolved in 50/zl TBS. Although this amplifies only 6% of the second eluate, this still represents the yield from at least 6000 TU/TU in the original unamplified first eluate. A 5-tzl portion of the amplified second eluate is then subjected to yet a third round of biopanning. The final (usually third) eluate is titered to give well-separated colonies (see titering transducing units), so that individual colonies can be propagated and analyzed (see Analysis of Affinity-Purified Phage, below).
Other Methods of Affinity Purification Two affinity matrices other than a polystyrene surface have been used for affinity purifying phage: agarose beads 1°,25 and oxirane acrylic particles. 11The agarose beads are large and have pores that are large compared to the ligate but small compared to the long dimension of the virion; it seems likely, therefore, that only a tiny fraction of immobilized ligate will be available to virions. The nature of binding to such an affinity matrix is not clear. Because the virions are presumably thin enough to partially penetrate the pores end first, for example, phage might bind a ligate just inside the interior and become "caged" in the immediate vicinity of the ligate; this could promote binding by increasing the effective off-time. The oxirane acrylic particles are small (-1/zm), hydrophilic, and nonporous; because they are not retained by any standard chromatography bed support, bound virions are separated from free virions by repeated centrifugation. Bound phages can be dissociated from immobilized ligand by means other than low pH, including competing free ligate or ligand and high pH (up to pH 11 for a few minutes). Phages are also resistant to urea (6 M for - 1 0 min, 4 M for >20 rain), sulfhydryl (e.g., 50 mM dithiothreitol), and trypsin (and probably many other specific proteases). Sometimes specific binding will be reversible, especially when the ligate is monovalent (see Discussion, below); in such circumstances binding will be manifested as delayed release even in a nondissociating wash buffer.
Quantifying Enrichment The progress of enrichment of binding clones can be followed crudely by spot-titering the input and output of the second and subsequent rounds 25 j. McCafferty, A. D. Griffiths, G. Winter, and D. J. Chiswell, Nature (London) 348, 552 (1990).
[15]
L I B R A R I E S O F P E P T I D E S D I S P L A Y E D BY P H A G E
249
of biopanning (see Titering Transducing Units, above). The input for each round is the amplified eluate from the previous round. Yields substantially in excess of 3 x 10-5%--the background yield with wild-type phage--indicate enrichment of binding clones, although the binding responsible for enrichment is not necessarily specific. Yields on the order of 1-5 x 10-4% are often associated with sequence motifs that recur regardless of ligate specificity, as will be discussed later. Yields higher than 0.01% usually (but not always) indicate the presence of a strong binding sequence motif in the phage-borne peptide. Seldom is the yield higher than - 1 % , even when ligate is in excess, and the individual clones that predominate in the input give higher yields when propagated and tested individually. Analysis of Affinity-Purified Phage Biopanning and other affinity purification techniques yield clones even when there is no specific binding of the ligate to phage-borne ligand-hence the need for independent confirmation of specific binding. The amino acid sequences of the displayed ligands (determined by sequencing the coding nucleotides in the viral DNA) are the primary and most informative data, but in many instances it is useful to supplement those data with direct binding assays, including ELISA and a miniaturized version of biopanning we call "micropanning." Analysis starts by processing small cultures of individual affinity-purified clones.
Sequencing The experimenter with many affinity-purified clones to analyze faces an unusual sequencing problem: the number of clones to process is large (we usually sequence 20 per experiment), but the number of unknown bases is small (e.g., 18 for a hexapeptide epitope library). Moreover, the penalty for mistakes is low, because little work has been invested in individual clones. The protocol described in General Procedures (see Sequencing Template, Sequencing Reactions, and Pouring and Running Sequencing Gels, above) attempts to meet this demand; its main features are (1) use of an end-labeled primer, eliminating the need for a separate "labeling" reaction for each template to be sequenced, (2) use of combinations of dideoxy terminators at different concentrations to reduce the number of lanes per template from four to two ("two-lane sequencing"), and (3) piggy-backing two sets of sequences on a single sequencing gel. The essence of two-lane sequencing is the use of termination mixes containing combinations of ddNTPs: the Q mix contains high concentrations of ddATP and ddCTP and a 10-fold lower concentration of ddCTP;
250
MUTAGENESIS::~vND GENE DISRUPTION
[15]
while the R mix contains high concentrations of ddATP and ddGTP and a 10-fold lower concentration of ddCTP. When Q and R reactions are run side by side in a sequencing gel, a strong band in both lanes signifies an A residue in the synthesized strand; a strong band in the Q lane (and no band in the R lane) signifies T; a strong band in the R lane (and no band in the Q lane) signifies G; and a weak band in both lanes signifies C. Because this method depends on reliable intensity differences, MnCI2 is substituted for MgCI2 in the buffer to reduce nonuniform incorporation of ddNTPs. 26 Two-lane sequencing works well for up to - 6 0 bases; beyond that point it becomes difficult to see the weak C bands over the entire sequence without multiple exposures of the X-ray film. For longer sequences we use four termination mixes with combinations of ddNTPs at a uniformly high concentration: mixture W has ddATP and ddTTP; M has ddATP and ddCTP; K has ddGTP and ddTTP; and S has ddGTP and ddCTP. The W, M, K, and S reactions are run in that order in the sequencing gel. Bands in both the W and M lanes signify A in the synthetic strand; bands in the W and K lanes signify T; bands in the M and S lanes signify C; and bands in the K and S lanes signify G. This system provides considerable redundancy of information (e.g., any three of the lanes contain enough information to deduce the sequence), and requires only side-by-side alignment of bands, thus greatly reducing ambiguity in ordering residues in the sequence. The concentrations of ddNTPs in the termination mixes are appropriate for sequencing up to - 6 0 bases from the primer. When longer inserts are to be sequenced, the termination mixes are diluted with termination diluent, which has no ddNTPs but contains all the other components of the termination mixes; a one-fourth dilution is suitable for sequencing up to -200 bases, and a one-eighth dilution for up to -400 bases.
ELISA and Other Binding Assays Binding of ligate to the phage-borne ligand can be demonstrated by ELISA; this serves to confirm binding in a way that is quite different from the affinity purification that isolated clones in the first place. Here we describe an ELISA in which phage themselves are immobilized in the wells of microtiter dishes. Phage are partially purified for ELISA, and 35-/zl portions containing - 2 × 101° virions are placed in microtiter wells overnight in the refrigerator; it is a good idea to include negative control phage prepared in the same way. The wells are filled with 100/xl Blotto solution (see Recipes, above), emptied by shaking out the solution and slapping the dish face down on a clean paper towel to remove 26 S. T a b o r and C. C. Richardson, Proc. Natl. Acad. Sci. U.S.A. 86, 4076 (1989).
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
251
residual solution, refilled with 350 /zl Blotto solution, and incubated at least 90 min at room temperature to block nonspecific binding sites on the plastic surface. The wells are emptied, washed three times with TBS/ Tween, and filled with biotinylated ligate in 35/xl TBS/Tween/azide containing 1 mg/ml BSA. Biotinylated monoclonal IgG and Fab fragments are typically used at concentrations of 1 and 300 nM, respectively; other ligates may require different concentrations. After reaction for 1-24 hr at - 4 °, the wells are emptied, washed once with TBS/Tween, filled with 85 t~l avidin-peroxidase complex (see Recipes, above), and incubated at room temperature for 15-30 min. The wells are emptied, thoroughly washed 10 times with TBS/Tween, and filled with 85/~1 freshly prepared ABTS solution (see Recipes, above). The wells are then allowed to react - 1 hr at room temperature before being read on a plate reader. The difference between A405 and A495 is taken as the signal, although A405 alone serves well enough. The strongest signals amount to -0.6, while background is -0.02. Only relatively strong ligands give signals in this assay. The foregoing method requires that phages be partially purified to reduce impurities that might interfere with adsorption to the plastic or give a high background signal. Purification is a tedious undertaking when hundreds of clones are to be assayed, and in many cases it can be circumvented with the aid of an anti-phage antibody directed against the major coat protein. Labeled anti-phage antibody can be used for specific detection of phages that have bound immobilized ligate25; alternatively, immobilized anti-phage antibody can be used for specific capture of phages in order to follow their reaction with labeled ligate] 7 An attractive alternative to binding assays that use the phages themselves is chemical synthesis of the phage-borne peptides on plastic pins in a microtiter array. 28 The pin-borne peptides can be used directly in a microtiter assay that confirms binding of the ligate to the peptides in an entirely independent way.
Micropanning Micropanning is a miniaturized version of biopanning designed to test affinity purification on many individual clones] We will give here a method appropriate for confirming tight binding of an antibody to phage, then discuss modifications for detecting weak binding. To each well of a microtiter dish (e.g., Immulon-2; Dynatech, Alexandria, VA) containing 20/~10.1 M NaHCO3 is added 1/x1200/zg/ml streptavi.,7 W. J. Dower, personal c o m m u n i c a t i o n , 1991. 28 H. M. G e y s e n , Immunol. Today 6, 364 (1985).
252
MUTAGENESIS AND GENE DISRUPTION
[15]
din; the streptavidin is allowed to coat the plastic overnight in a humidified chamber in the refrigerator. Meanwhile, the wells of another microtiter dish are filled with 20/zl 0.1-1 nM biotinylated antibody in TBS/azide with 1 mg/ml dialyzed BSA, and 1/xl of each culture supernatant (typically 4.5 × l 0 7 T U ) is added to one of the wells; the reaction is allowed to proceed overnight in the refrigerator in a humidified chamber. The next day the wells of the streptavidin-coated plate are emptied, filled with blocking solution, reacted 1 hr at room temperature, and emptied and washed four times with TBS/Tween. Meanwhile, the phage-antibody reactions are diluted with 90/zl TBS/Tween and 15-/zl portions are transferred into the streptavidin-coated wells. After I0 min at room temperature the wells are emptied, washed I0 times with TBS/Tween, and eluted with 20/xl elution buffer. After 10 min at room temperature the elution buffer is neutralized by adding 140 /xl of a 69:1 (v/v) mixture of TBS and 2 M Tris (pH unadjusted). Four-microliter portions of the neutralized eluates are pipetted into the wells of a microtiter dish containing 10/~1 K-91 starved cells, and after 10 min at room temperature 336 ~1NZY medium containing 0.2/zg/ml Tc is added to each well. After incubation at 37 ° for 30-45 min, 20-/xl portions are spotted onto NZY/Tc agar medium. Strongly positive clones give >50 colonies per spot, indicating a yield of - 1 % , while negative clones give 0-1 colony per spot, possibly representing contamination from neighboring wells rather than genuine yield. This procedure can be modified at several points to detect weaker binding: (1) the phage-antibody reactions need not be diluted before the 15-/.,1 portions are pipetted into the streptavidin-coated dish; (2) eluates can be neutralized with 1.2 ~12 M Tris; (3) 20-~1 portions of the neutralized eluates can be spotted directly on an NZY/Tc agar medium that has previously been seeded with 1 ml of log-phase K-91 culture poured in - 3 ml soft agar. When weak binding is to be detected, it is also advantageous to titer dilutions of the original culture supernatants to control for variations in phage yield.
Summary of Results In this section we emphasize experiments with our hexapeptide epitope library, in which the N-terminal sequence of pIII is ADGAX6GAAGA-, X6 being the variable peptide. We have biopanned the library with at least five monoclonal antibodies (MAbs) that were known or suspected in advance to be specific for continuous peptide epitopes--that is, epitopes composed of amino acids that are contiguous in the primary sequence of the eliciting antigen. All of them identified peptide ligands with a Consensus sequence similar to that of the
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
253
eliciting epitope. Often there were a few differences between the phageborne peptides and the eliciting epitope; but in those cases in which the specificity of the MAb had been studied independently by binding studies with synthetic peptides, the differences turned out to be mostly at positions that are tolerant of multiple substitutions. When binding of the antibody to phage has been checked by ELISA, strong binding is evident (ELISA signal, 0.3-0.6). The foregoing results contrast with those using ligates that were not known in advance to bind continuous epitopes. In Table II we summarize results of three parallel biopannings with such ligates. MAb 3-4B is a mouse monoclonal IgM that reacts with the large subunit of Caenorhabditis elegans RNA polymerase II and with a fusion protein containing 148 residues of that polypeptide] 9 Insulin normally binds a cell surface receptor, although it has also been shown to bind a hexapeptide. 3° IL-2R~ is the extracellular binding domain of the small (o0 subunit of the IL-2 receptor; it binds the polypeptide hormone IL-2 without the large subunit. 31'32 For each ligate the table shows the peptide sequences (X6) displayed by each affinity-purified clone, and the ELISA signal with each of the three ligates. As Table II shows, a single peptide sequence PTWRSM predominates among the clones affinity purified with MAb 3-4B. Phage displaying this peptide reacted weakly with the MAb (ELISA signal, 0.03), but still at least 10-fold above the background signal and all the other ELISA signals reported in Table II (0.000-0.003). The consensus peptide shows no convincing similarity to any continuous hexapeptide segment of the antigen. This contrasts with our results using antibodies specific for continuous epitopes discussed above, and raises the possibility that PTWRSM is an example of what Geysen and coUeagues 33'34 call a " m i m o t o p e " m a short peptide that mimicks the binding properties of an assembled epitope (one formed from residues distant in the primary sequence of the antigen but contiguous in its folded structure). A few other MAbs specific for assembled epitopes have identified putative mimotopes in our epitope library, ~9 j. Prenger, M. Golomb, and J. Jones, personal communication (1991). 30 V. P. Knutson, J. Biol. Chem. 263, 14146 (1988). ~ B. F. Treiger, W. J. Leonard, P. Svetlik, L. A. Rubin, D. L. Nelson, and W. C. Greene, J. lmmunol. 136, 4099 (1986). 32 j. Haikimi, C. Seals, L. E. Anderson, F. J. Podlaski, P. Lin, W. Danho, J. C. Jenson, A. Perkins, P. E. Donadio, P. C. Familletti, Y. E. Pan, W. Tsien, R. A. Chizzonite, L. Casabo, D. L. Nelson, and B. R. Cullen, J. Biol. Chem. 36, 17336 (1987). 33 H. M. Geysen, S. M. Rodda, and T. J. Mason, in "Synthetic Peptides as Antigens" (R. Porter and J. Whelan, eds.), p. 130. Wiley, New York, 1986. 34 H. M. Geysen, S. M. Rodda, and T. J. Mason, Mol. Immunol. 23, 709 (1986).
254
[15]
MUTAGENESIS AND GENE DISRUPTION T A B L E 1I RESULTS OF BIOPANNING EXPERIMENTa
E L I S A signal ( × 103) with Ligate
N u m b e r of clones
MAb3-4B
10 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1
Insulin
IL-2Ra
X 6 sequence
PTWRSM TRMRPG VLLSVA INQVRF WCSRLF PCHCSF RGYFFK ATWAVL AVMTSS SWFLQW REWlSH YTALLI CYLCSV LFSSGK VWHLLH VPWWVP LSRILF PGHSPW WNLRSS IALMDY RAWSYV KGRYQQ EHGRPQ GCSDVL NLLSMT CLGEHD RFYGGS ILPLRI LSRSYF NLYLVH AWFRRL LRGKLS VDVGRS
M A b 3-4b
Insulin
IL-2Ra
30-33 2 1 2 2 1 0 0 0 0 0 0 - 1 1 3 0 1 1 0 0 0 1 0 - 1 - I 0 - 1 1 0 2 0 - 1 1
0-3 3 3 3 3 1 2 I 1 1 1 1 0 4 0 1 0 0 1 1 0 0 0 1 0 1 0 1 1 0 0 1 1
0-3 3 2 3 2 1 1 1 1 1 1 1 1 2 1 0 1 1 1 1 2 1 2 2 1 1 1 1 1 1 1 1 2
a The experiment is explained in text.
but most such antibodies have given results similar to those with insulin and IL-2Ra to be discussed in the next paragraph. Unlike MAb 3-4B, neither insulin nor IL-2Ra yielded affinity-purified clones with a consensus sequence; nor did any of the phages bind either ligate by ELISA (Table II). At best, therefore, these phages display weak ligands with little sequence specificity; and it is entirely possible that they
[15]
L I B R A R I E S O F P E P T I D E S D I S P L A Y E D BY P H A G E
255
represent merely the low background of nonspecific yield observed when any nonbinding phage is subjected to the biopanning procedure ( - 3 x 10-5%). These data are included here merely as an example of sequence and ELISA results we ordinarily regard as negative, and should not be overinterpreted. For instance, we did not show by any independent experiment that either biotinylated ligate retained its binding capacity. Furthermore, unlike antibodies, both these ligates are monovalent and for that reason may release even tight-binding phages during the long washes of the biopanning procedure (see Discussion, below). Two consensus sequences are observed repeatedly with a variety of antibodies and other ligates: PWflWLX (where fl is usually either A or E), and GDWVFI and related variants. Presumably, these peptides bind some component of the system other than the ligate. The binding must be weak, because the peptides never appear when strong ligands are identified with antibodies against continuous epitopes. The presence of these peptide sequences indicates that the ligate has not identified a strong ligand in the library. Although the absence of a strong ligand is presumably a necessary condition for observing these peptides, it probably is not a sufficient condition: if none of these weak ligands happens to survive the first round of biopanning, when each clone is represented by only -100 TU, it obviously cannot become a consensus sequence after further rounds. Discussion
Biopanning and Kinetics of Binding The mechanism of purifying binding phages from the library is complex. Phages are reacted with a biotinylated ligate, then the complexes are diluted, bound to streptavidin-coated plates, washed extensively, and eluted. The apparent off-rate of phage-borne peptides must be low for them to remain bound through 30-60 rain of washing. In cases where dissociation constants of free peptides were known, the off-rates of phages bearing a peptide of known sequence were orders of magnitude lower than the presumed off-rate of the corresponding free peptide. This discrepancy may be explained by assuming that the phages, which bear up to five copies of each epitope, allow multivalent binding to the antigen-binding sites on antibodies (and some receptors). This "avidity boost" would greatly reduce the apparent binding off-rate, because once binding has occurred there would be a much lower chance of both binding sites being released simultaneously. It is also possible that other factors, such as the slower diffusion rate of phage compared with free peptide, may play a role in this effect. In every case tested, when intact, bivalent antibody contin-
256
MUTAGENESIS AND GENE DISRUPTION
[15]
ues to bind phage after extensive washing, the corresponding Fab' does also; however, about a 100-fold higher concentration of Fab' is required to produce the same effect. Although this may reflect true monovalent affinity of the binding site for phage-borne ligand, it is equally possible that it is due to the presence of a small fraction of bivalent species in the Fab' preparations. If multivalency of binding sites is required for biopanning to be successful, it is not surprising that, as yet, even moderateaffinity peptides have not been identified when the hexapeptide library has been screened with monovalent ligates such as insulin and IL-2Ra (see above). The effect of valency on binding off-rate awaits testing with indisputably monovalent ligates or phage-borne epitopes.
Can Ligands Be Found for Most Binding Proteins? In the beginning we hoped the hexapeptide epitope library would contain strong ligands for virtually any binding protein. In the case of antibodies against continuous peptide epitopes this hope has been fulfilled, but as discussed above there are many ligates--particularly antibodies against assembled epitopes--for which tight ligands cannot be isolated from the hexapeptide epitope library. There are at least two possible reasons for this failure. First, some binding sites may require structural features that cannot be effectively mimicked by short peptides. Second, the particular invariant residues that flank the hexapeptide epitope in our library may be nonpermissive for binding to some ligates. These difficulties might be overcome with epitope libraries that display longer variable regions, or with "constrained" epitope libraries in which structural constraints restrict the variable residues to particular conformations. Such constraints might be imposed by disulfide bonds, or by displaying the variable residues in the context of a small structured domain fused to the coat protein (see the next section). We speculate that a limited number of structured libraries might suffice to provide strong ligands for almost any ligate.
Cloning Structured Domains A variety of small protein domains have been fused to pIII and pVIII for the purposes of making libraries and studying mutant proteins. Bass and co-workers H expressed human growth hormone (hGH) on virions using a construct in which the exposed N-terminal domain of pIII was replaced with the 191-residue hormone. The hGH-phage retained full infectivity, as no more than one of the five pIII proteins on each virion was recombinant; the remaining pIII molecules were supplied by a helper phage, and were wild type. The hGH-phage bound hGH receptor and MAbs that recognize only the native structure, showing that the fusion
[15]
LIBRARIES OF PEPTIDES DISPLAYED BY PHAGE
257
protein folded properly, and could be specifically enriched from nonrecombinant phage on hGH receptor-coated beads. McCafferty et al. z5 constructed an N-terminal pIIl fusion displaying a single-chain antibody specific for hen egg-white lysozyme. These so-called "phage antibodies" retained native specificity and could be enriched from nonrecombinant phage on lysozyme affinity columns. Parallel work by Kang et al. 12 shows that as many as 30 Fab domains can be displayed along the length of virions as fusions to the gene VIII protein; the bulk of the pVIII molecules are wild-type subunits encoded by a helper phage. These examples demonstrate that foreign domains displayed by phage can retain at least partial native folding and activity, and that phage displaying these fusion products can be selectively enriched by affinity purification. This capability suggests the feasibility of libraries in which sequence variability (ranging from a low frequency of substitutions to totally random sequences) can be targeted to specific regions in a folded protein domain. P h a g e - A n t i b o d y Libraries
Phage-antibody libraries may be seen as an extreme form of constrained epitope library. The immunoglobulin framework residues provide a rigid scaffolding for displaying six variable peptides--the so-called complementarity-determining regions (CDRs), three of which are present in both the light and heavy chain variable regions. It is the CDRs that primarily determine the binding specificity of an antibody. Different clones in the phage-antibody library would display different antibody domains with different specificities. A great diversity of specificities would be incorporated into the library as a whole, either by cloning the natural repertoire of antibody genes present in animals, or by randomizing the CDRs. An ideal phage-antibody library--not yet a reality--would include antibodies specific for any antigen. It would be manufactured on a large scale and distributed to multiple experimenters, who would use their chosen antigens to affinity purify out of the library those phage whose displayed antibody domains bind with highest affinity. Affinity might be improved by subjecting these initial clones to further rounds of random mutagenesis and selection by affinity purification. These may be the monoclonal antibodies of the future: produced without need for animals or animal cells in culture, and available to any laboratory able to carry out simple recombinant DNA techniques.
258
MUTAGENESIS AND GENE DISRUPTION
[16]
[16] U s e o f M 1 3 P i n g - P o n g V e c t o r s a n d T 4 D N A P o l y m e r a s e in O l i g o d e o x y n u c l e o t i d e - D i r e c t e d M u t a g e n e s i s B y MARY M . Y . W A V E
Introduction Oligodeoxynucleotide-directed mutagenesis has undergone rapid improvement since its introduction. 1 Although "single priming" without strand selection using a repair host strain generally gives acceptable resuits, 2,3 various methods of improving the efficiency of mutagenesis have been developed. They rely on selection against the parental template, via (1) the use of EcoK/EcoB selection, 2'4 (2) the use of nonsense c o d o n s Y (3) the use of hemimethylated D N A , 6'7 (4) the use of phosphorothioatemodified DNA, 8-1° or (5) the use of a uracil-containing parental template that is inactivated by uracil N-glycosylase. 11.12Deletion mutagenesis using oligodeoxynucleotides can be problematic without any selection. The frequency of deletion mutants can vary from 0.4 to 2 0 % . 4'13 This chapter describes a new vector, M13B119, and an improved strategy for "cyclic selection" that is especially useful for creating multiple mutations with large deletions. Another new vector, M13K119W, is also used to demonstrate how the conditions for oligodeoxynucleotide-directed mutagenesis can be optimized. For example, an improvement of 2- to 18-fold can be obtained with deletional mutagenesis using T4 DNA polymerase. However, the use of different polymerases, including T4 polymerase, Sequenase (trade name for a modified T7 DNA polymerase; United States Biot M, Smith, Annu. Rev. Genet. 19, 423 (1985). 2 p. Carter, H. Bedouelle, and G. Winter, Nucleic Acids Res. 13, 4431 (1985). 3 p. Carter, this series, Vol. 154, p. 382. 4 M. M. Y. Waye, M. E. Verhoeyen, P. T. Jones, and G. Winter, Nucleic Acids Res. 13, 8561 (1985). 5 W. Kramer, V. Drutsa, H. W. Jansen, B. Kramer, M. Pflugfelder, and H.-J. Fritz, Nucleic Acids Res. 12, 9441 (1984). 6 W. Kramer, K. Schughart, and H.-J. Fritz, Nucleic Acids Res. 10, 6475 (1982). 7 A. Marmenout, E. Remaut, J. Van Boom, and W, Fiers, Mol. Gen. Genet. 195, 126 (1984). 8 D. B. Olsen, J. R. Sayers, and F. Eckstein, this volume [13]. 9 j. W. Taylor, W. Schmidt, R. Cosstick, A. Okruszek, and F. Eckstein, Nucleic Acids Res. 13, 8749 (1985), 10j. W. Taylor, J. Ott, and F. Eckstein, Nucleic Acids Res. 13, 8765 (1985). ii T. A. Kunkel, Proc. Natl. Acad. Sci. U.S.A. 82, 488 (1985). lZ T. A. Kunkel, J. D. Roberts, and R. A. Zakour, this series, Vol. 154, p. 367. 13 V.-L. Chan and M. Smith, Nucleic Acids Res. 12, 2407 (1984).
METHODS IN ENZYMOLOGY.VOL. 217
Copyright© 1993by AcademicPress, Inc. All fightsof reproductionin any formreserved.
[16]
M13 VECTORS FOR SITE-DIRECTED MUTAGENESIS
259
chemical Corporation, Cleveland, OH), and pollk (the Klenow fragment of Escherichia coli DNA polymerase), gives similar ratios of true mutants to spurious deletional mutants. The use of higher temperature (37 or 20° instead of 15°) helps to reduce the percentage of spurious deletional mutants by approximately 10-20%.
Principles of Method
Design of Oligonucleotide Primers The design of oligonucleotide primers has been reviewed, z Competing priming sites can be avoided by comparing the proposed sequence of the oligonucleotide with that of the vector and cloned insert by using a computer program such as ANALYSEQ. ~4 Furthermore, it is necessary to ensure that the two primers do not have large regions of homology when using "coupled primer" selection. More recently, computer programs, for example, Primer Designer, GeneJockey, and Oligo have also been developed to evaluate primer designs. Primer Designer is available from Scientific and Educational Software (PA), GeneJockey is available from Biosoft (Cambridge, England), and Oligo is available from National Biosciences (Plymouth, MN).
Use of EcoK/EcoB Selection If double-stranded DNA containing an EcoK site is introduced into a K strain (rK + mK+), the DNA will be either modified if one strand is already modified or it will be restricted if neither strand is modified. ~5-17The same principles apply to EcoB sites introduced into a B strain (rB + mB+). 18 Because the M13 DNA replicates rapidly and the amount of EcoK enzyme in a bacterium is limiting, a substantial amount of DNA escapes restriction if only one copy of EcoK is used as a selection site. This problem can be overcome by using four tandem copies of the EcoK site as the selection site. 4 This chapter describes the use of new vectors with four overlapping copies of EcoB or EcoK sites so that the selection primer and the selection site can be smaller and more portable. 14 R. Staden, Nucleic Acids Res. 12, 521 (1984). 15 N. C. Kan, J. A. Lautenberger, M. H. Edgell, and C. A. Hutchison IIl, J. Mol. Biol. 130, 191 (1979). 16 p. Modrich, Q. Rev. Biophys. 12, 315 (1979). 17 R. Yuan, Annu. Rev. Biochem. 50, 285 (1981). i8 j. A. Lautenberge, M. H. Edgell, C. A. Hutchison III, and G. N. Godson, J. Mol. Biol. 131, 871 (1979).
260
MUTAGENESIS AND GENE DISRUPTION
[16]
Use of M13 Ping-Pong Vectors The technique described in this chapter is designed to increase mutant yield by selecting against progeny phage from the parental M13. One oligonucleotide is used to construct the "silent" deletion or mutation of interest and a second primer (selection primer 1) to remove a selectable marker (four overlapping copies of EcoB) in the template. These two primers are then extended by T4 DNA polymerase and the nascent strands ligated to the kinased primers. The heteroduplex DNA is then used to transfect an E. coli host strain that has the EcoB restriction enzyme so that the template strands and any progeny that copied the EcoB sites will be selected against. Only the progeny phage from which the selectable marker (four overlapping copies of EcoB) has been removed will be viable. At the same time that the first selection marker (four overlapping copies of EcoB) is removed, another selection marker (four overlapping copies of EcoK) is introduced by the first selection primer so that the process can be repeated again for constructing a second, "silent" deletion or mutation of interest. The second silent deletion will be accompanied by a second selection primer 2, which will remove a selectable marker (four overlapping copies of EcoK) from the template and add another selection marker (four overlapping copies of EcoB). Thus, multiple deletion or mutations can be constructed by cycling between these two selectable markers.
Use of Different DNA Polymerases The three different polymerases used in this study have different properties. T4 DNA polymerase has a 5' ~ 3' polymerizing activity and a 3' ~ 5'-exonuclease activity; unlike E. coli DNA polymerase, it does not have a 5' ~ 3'-exonuclease activity.19 The Klenow fragment of E. coli DNA polymerase I has a 5' ~ 3' polymerizing activity and a 3' --~ 5'-exonuclease activity, but lacks the 5' ~ 3'-exonuclease activity, z° Sequenase, which is a modified T7 DNA polymerase, has high processivity and low 3' ~ 5'-exonuclease activity. 2~ T4 DNA polymerase has been used for oligodeoxynucleotide-directed mutagenesis in preference to the Klenow fragment of E. coli DNA polymerase 1222due to its lack of strand displacement of the mutagenic primer, z3 This chapter describes the use of the EcoK/EcoB vectors in choosing between the three polymerases and in optimizing the temperature of the polymerization/ligation reaction. 19 N. G. Nossal, J. Biol. Chem. 249, 5668 (1974). 20 H. Jacobsen, H. Klenow, and K. Overgaard-Hansen, Eur. J. Biochem. 45, 623 (1974). 2l S. Tabor and C. C. Richardson, Proc. Natl. Acad. Sci. U.S.A. 84, 4767 (1987). 22 j. Geisselsoder, F. Witney, and P. Yuckenberg, BioTechniques 5, 785 (1987). 23 M. Goulian, Z. J. Lucas, and A. Kornberg, J. Biol. Chem. 243, 627 (1968).
[16]
M13
VECTORS
FOR
261
SITE-DIRECTED MUTAGENESIS
TABLE I M13 VECTORS WITH EcoK/EcoB SELECTION MARKER(S) MI3 strains MI3KI8 M13K19 M13BI8 M13BI9 M13Kll MI3K11RX M13KII9 M13BI19W MI3KII9W MI3BlI9
EcoK
EcoB
Blue/White
EcoB in II a
Ref.
---
B B
---
b b
1x
B
--
b
1x
-+ + -----
b
---4 × ovlp -4 x ovlp
B B B B W W B
1x 1x
4x 4x 4x 4x
--tdm" tdm tdm -ovlp" --
c d d d, e d, f
" EcoB in II: the presence of EcoB site in gene H of M 13; tdm, tandem; ovlp, overlapping. b p. Carter, H. Bedouelle, and G. Winter, Nucleic Acids Res. 13, 4431 (1985). ' M. M. Y. Waye, M. E. Verhoeyen, P. T. Jones, and G. Winter, Nucleic Acids Res. 13, 8561 (1985). d M. M. Y. Waye, F. Mui, and K. Wong, Technique 1, 188 (1989). e Starting clone for studies of different polymerases and polymerization/ligation temperature. Starting vector used for c-fos cyclic deletion experiments described as an example in this chapter. Materials
and Methods
Experimental Details Enzymes. T4 DNA polymerase and T4 DNA ligase are purchased from P h a r m a c i a ( P i s c a t a w a y , N J ) ; p o l I k , t h e K l e n o w f r a g m e n t o f E. coli D N A p o l y m e r a s e I, is p u r c h a s e d f r o m e i t h e r B e t h e s d a R e s e a r c h L a b o r a t o r i e s ( G a i t h e r s b u r g , M D ) o r P h a r m a c i a ; S e q u e n a s e is p u r c h a s e d f r o m U S B (United States Biochemical Corporation). B a c t e r i a l S t r a i n s a n d V e c t o r s . T h e f o l l o w i n g E. coli s t r a i n s a r e u s e d : A C 2 5 2 2 : A n E. coli s t r a i n t h a t h a s E c o B b u t n o E c o K r e s t r i c t i o n e n z y m e : B / r , H f r , sul-124 J M 1 0 1 : a n E. coli s t r a i n t h a t h a s E c o K b u t n o E c o B r e s t r i c t i o n e n z y m e 25 T G I : A n E. coli K - 1 2 s t r a i n t h a t h a s n o E c o K r e s t r i c t i o n e n z y m e 26 Vector M13K119W: Used to test the different conditions Vector M13Bl19: Used as the starting vector for the cyclic deletion e x p e r i m e n t s ( s e e T a b l e I) 24 H. Boyer, J. Bacteriol. 91, 1767 (1966). -,5 j. Messing, Recomb. DNA Tech. Bull. 2, 43 (1979). 26 T. J. Gibson, Ph.D. thesis. University of Cambridge, Cambridge, England, 1984.
262
MUTAGENESIS AND GENE DISRUPTION
[16]
Methods for mutagenesis are essentially the same as those described previously, 2'27'28 except for the following:
Hybridization 1. Anneal primer and template together in an Eppendorf tube: Kinased primer 1 (5 pmol//zl), 2/zl Kinased primer 2 (5 pmol//zl), 2 ~1 Template (1 /~g/~l), 1 /xl TM buffer (10 × ) [100 mM Tris-HC1 (pH 8.0), 100 mM MgClz], 1/zl Water, 4/zl 2. Place the tube containing the sample in a small beaker of hot water (80°), and let cool to room temperature. This will take about 30 min.
Extension/Ligation I. Add to the annealing mix: TM Buffer (I0 x ), 1 p.l rATP (5 mM), 1 /xl dNTPs (5 mM), 1 /xl Dithiothreitol (DTT) (100 mM), 1 p.l Water, 4/xl 2. Place on ice and then add T4 DNA ligase (10 U; Pharmacia) and Klenow fragment of DNA polymerase (3 U; Pharmacia). Then incubate 12-20 hr at 15° or other specified temperatures. For experiments on testing different polymerases, either Sequenase (3 U) or T4 DNA polymerase (3 U) was used instead of the Klenow fragment of DNA Polymerase, and the temperature of extension/ligation was either 15, 20, or 37°.
Results of Experiment The four times-overlapping EcoB vector was used to delete all three introns of the c-fos gene by a double-primer strategy (see Fig. 1 for a schematic diagram, Fig. 2 for the DNA sequence of the selection primers, and Fig. 3 for the polylinker region of the vector M13B 119). The following oligonucleotides were used: The 45-mer (sel 1): 5'-CCC TA(G CAC GCA CCG GTT AGT TGC ACG CAC ACG TTA GTT) TCA TTG-3' was used to construct M13K119. The nucleotides in parentheses are complementary to nucleotides that 27 W.-Y. Shen and M. M. Y. Waye, this series, Vol. 218, pp. 58-71. 28 W. Shen and M. M. Y. Waye, Gene 70, 205 (1988).
[16]
263
M13 VECTORS FOR SITE-DIRECTEDMUTAGENESIS 4X ECOB
Grow M13 c-fos in TG1 (EcoK," Ecol')
B
+ DEL #1 PRIMER + SEL #1 PRIMER (4 X ECOK ) M13 C-FOS
AC2522
ECOWECOB SELECTION
( ECOB +)
+ SEL#2 PRIMER (4 X ECOB) + DEL #3 PRIMER
B
SELECT IN JM101 ( ECOK + )
0
+ SEL #1 PRIMER ( 4 X ECOK)
K L.
+ DEL #2 PRIMER l
(9 SELECT IN AC2522 ( ECOB +)
K
i3cFosl intron -
FIG. 1. A schematic diagram of the cyclic deletion strategy using double primers and the MI3Bl19 vector. Intron 1 was removed first, followed by removal of introns 3 and 2.
h a v e four overlapping copies of the EcoK site. The deletion primer for deleting the introns were (deI1):5'-dGTCAACGCGCAGGACTTCTGCACG (deI2):5'-dAAGGTGGAACAGTTATCTCCAGAA (del3):5'-dACACTCCAAGCGGAGACAGACCAA The sequencing p r i m e r (spl) for analyzing the deletion in the first intron was 5 ' - d T C C C G T T G T G A A G A C C A .
Cloning of c-fos D N A Insert into an M13 Ping-Pong Vector M13B119. The plasmid pF4 (which has the 5.4-kb BamHI fragment of the h u m a n
264
MUTAGENESIS AND GENE DISRUPTION l E(X~ I
[16]
TGN~NNNNNNNTGCT TGANNNNNNNNTGCT TGANNNNNNNNTGCT
I
T j GANNNNNNNNTC.-,CT
M13B119 IA'I-rCCCAACCTGAAACTGACGTGCTGATGCTGATGTGCTCGTGC'I-rAAGGGGATCC 3' 56-MER ITAAGGGTTGGACTI-I'GACTGCACGACTACGACTACACGAGCACGAATTCCCCTAGG 5' # 2 51-MER Ir CTTAAGGG'I-I'ACTI-rGACTGCACGACTACGACTACACGAGCACGATCCCCT. 5' #1 45-MER & 3 ' GI'I'AC'ITI'GA'I-I"GCACACGCACGI-I'GA1-FGGCCACGCACGATCCC5' M13K119W [ 5' CAATGAAACTAACGTGTGCGTGCAACTAACCGGTGCGTGCTA~3' AACNNNNN~TGC E(X]K AACNNNNNNGTGC l AACNNNNNNGTGC
I
I'
MC,NNNNNNGT~ FIG. 2. The DNA sequence of the selection primers 1 and 2 used for the cyclic deletion strategy described in Fig. 1 and the 56-mer oligonucleotide used for constructing the vector M I3B 119. The overlapping EcoK and EcoB sites are shown below and above the primers, respectively.
c-los gene 29'3°) was obtained from N. Miyamoto (Ontario Cancer Institute, Toronto, Canada). The AurlI-NotI fragment (2.657 kb) ofpF4 was cloned into the EcoRV site of vector M 13B 119 and the orientation was confirmed by D N A sequencing using the universal priming site. The resultant clone was named M13 c-fos and single-stranded D N A template was grown in TG1 (EcoK-, EcoB-). Removal of First Intron. Both the selection primer (sel 1, a 45-mer with four copies of EcoK) and the deletion primer (del 1) were used for the mutagenesis experiment and the heteroduplex was transformed into AC2522 (EcoB +). When selection was used (by transfecting AC2522 with the heteroduplex), 28% of the transformants obtained were hybridization positive with the deletion primer and 30% of the clones sequenced were true positive. (Thus an overall percentage true positive of 28% x 30% = 8%.) However, when no selection was applied (by using either TG1 cells or by omitting the selection oligonucleotide, only 1 or 0% of overall true positive clones were obtained (see Table II and Table III). Removal of Third Intron. Both the selection primer (sel 2, a 51-mer with four copies of EcoB) and the deletion primer (del 3) were used for the mutagenesis experiment and the heteroduplex was transformed into JM101 (EcoK+). 29 T. Curran, W. P. MacConnell, F. Van Straaten, and I. M. Verma, Mol. Cell. Biol. 3, 914 (1983). 3o F. Van Straaten, R. MOiler, T. Curran, C. Van Veveren, and I. M. Verma, Proc. Natl. Acad. Sci. U.S.A. 80, 3183 (1983).
[16]
M13 VECTORS FOR SITE-DIRECTED MUTAGENESIS
lacZ ~ M T M I T ATGACCATGATT~ TACIWoG~ACTAA -
265
AccI SalI Hind II
Hind Ill
TAC~L~GATATCGAG TCTCL~ATAGCI~ Xbal EcoRV
PstI
Four EcoB sites . o . °
° . o °
° o o °
. . ° °
::: ::: ::: ::: GGTAGGAA~CmAACTGA~~YGIGCTCGTCf2AAA CF_ATCCI~~CITIW_~~CTACGACTACACGA~_ACV_~ATTT ECORI
FXa recognition site 4V---R G E I GGGC~TCCGTCGACCTA~TGGA _ T'IV_A~ CCCL~AGGCA~T C~GAGCTACCTAGGGC4X~~AAGSC~EF/;G - -~I SacI - Bar~qI SalI AccI BamHI EcoRI SmaI FspI
HgiEII
// PvuI
Bgl
Ystll
universal sequencing primer
FIG. 3. The DNA sequence of the polylinker region of the vector MI3B119. The colons above the sequence indicate the first 3 bp of the EcoB restriction sites. The dots above the sequence indicate the last 4 bp of the EcoB restriction sites.
Removal of Second Intron. Both the selection primer (sel l, a 45-mer with four copies of EcoK) and the deletion primer (del 2) were used for the mutagenesis experiment and the heteroduplex was transformed into AC2522, which is EcoB +. Conditions for the mutagenesis experiment used in deleting the introns of c-los were as previously described, 4 except that two primers were used in this study and the starting vectors were different. Test of different DNA Polymerases and Temperature of Extension/ Ligation. The phage M 1 3 K l l 9 W , which has four copies of EcoK and a
266
[16]
MUTAGENESIS AND GENE DISRUPTION
T A B L E II EFFECT OF E c o K OR E c o B SELECTION ON FREQUENCY OF HYBRIDIZATION POSITIVES a Oligo used
Results obtained with hybridization E. coli used
Deletion oligo
Selection oligo
A1 A1 A1
N u m b e r of base pairs deleted
EcoK/B
Strain
45-mer 45-mer None
EcoB +
AC2522 TG 1 TG1
753
A2 A2 A2
45-mer 45-mer None
EcoB ÷
AC2522 TG1 TG1
431
A3 A3 A3
51-mer 51-mer None
EcoK +
JM101 TG1 TGI
113
---
---
---
Number positive
Number screened
Percentage positive
Improved ratio
27 16 32
97 200 200
28 8 16
23x
65 38 37
197 200 200
33 19 19
1.7x
35 13 15
85 88 88
41 15 17
2.6x
Oligo, oligodeoxynucleotide. Improved ratios were obtained by dividing the percentrage positive of the first row by the average percentage positive of the controls (second and third rows) for each group.
white phenotype, was used for the experiments described in Figs. 4 and 5 of this study. The selection primer (a 56-mer oligonucleotide) was used to generate M13B 119, which has four overlapping copies of EcoB, using M13K119W as the template. Mutants that have lost the EcoK sites were selected by transfecting the heteroplex into JM101 cells, which have the EcoK restriction enzyme. The 56-mer deleted the four copies of EcoK while introducing the four copies of EcoB. The use of higher temperatures (37 or 20 ° instead of 15°) increases the efficiency of mutagenesis T4 DNA polymerase by 10-20% (see Fig. 4, compare bars on the left or bars on the right). In the presence of selection (i.e., by transfecting the heteroduplex in JM101), Sequenase and T4 DNA polymerase or pollk do not differ significantly in their ability to generate T A B L E III EFFECT OF E c o K OR E c o B SELECTION ON FREQUENCY OF CLONES WITH CORRECT SEQUENCE Results obtained by sequencing
Oligo used E. coli used
Deletion oligo
Selection oligo
AI A1 AI
45-met 45-mer None
EcoK/B
Strain
True positives
Spurious mutants
EcoB
AC2522 TGI TG1
3 1 0
10 7 10
---
Percentage true positives (3/10), 30% (1/7), 14% (0/10), 0%
Overall percentage true positives a (28% x 30%) = 8% (8% × 14%) = 1% (16% × 0%) = 0%
a Oligo, oligodeoxynucleotide. Overall percentage true positives are obtained by multiplying the percentage hybridization positives from Table II by the percentage true (sequenced) positives from Table III.
[16]
M I 3 VECTORS FOR SITE-DIRECTED MUTAGENESIS
267
80
6O
4o
20
0 3.25
6.25
Units of Enzyme used FIG. 4. The difference in the efficiency of mutagenesis obtained at different temperatures of incubation. Black bars, T4, 37°, JM + ; white bars, T4, 20°, JM + ; gray bars, T4, 15°, JM +.
nonspurious deletional mutants (compare the black bars in Fig. 5); in the absence of selection, T4 polymerase gives higher percentages of mutants (compare the open bars in Fig. 5). DNA sequence analysis was performed to confirm that blue plaques are bona fide mutants: 24 of 24 clones sequenced have the correct sequence, whereas white plaques have either 80
60
40
~-
20
T4 pol
Pollk
Sequenase
E n z y m e used FIG. 5. The effect of using different enzymes on the frequency of mutation, n , With selection; [], without selection. T4 pol, T4 DNA polymerase; Pollk, Klenow fragment of E. co/i DNA polymerase; Sequenase (trade name for modified T7 DNA polymerase).
268
MUTAGENESIS AND GENE DISRUPTION
[16]
shifted reading frames (15 out of 24 sequenced) or their universal priming site deleted (9 out of 24 sequenced).
Schemes 1. Synthesize the following selection primers: The 45-mer (sel 1): 5'-CCC (TA(G CAC GCA CCG GTT AGT TGC ACG CAC ACG TTA GTT) TCA TTG-3' and the 51-mer (sel 2): 5'-TCC CCT (AGC ACG AGC ACA TCA GCA TCA GCA CGT CAG TTT CA)T TGG GAA TTC-3'. Purify the primers by polyacrylamide gel electrophoresis. 2. Design and synthesize the appropriate mutagenic and sequencing oligonucleotides for analysis of the deletion junction. 3. Grow (e.g., 50/zg from 100-ml cultures) RF of M13Bl19 in TG1 cells, digest the vector DNA and also the DNA of interest, then clone the insert into any unique restriction site upstream of the EcoB selection sites of M13Bl19 (e.g., the EcoRV site). 4. Grow single-strand DNA template of the recombinant M13 in TG1 for mutagenesis (e.g., 18/xg from 6-ml cultures). 5. Kinase the deletion primer and selection primer 1. Anneal the recombinant phage DNA with the two primers, extend using T4 DNA polymerase, and ligate with T4 DNA ligase at 37°. Prepare overnight cultures of AC2522. 6. Transform AC2522 with the heteroduplex using the method of Hanahan. 3~Pick plaques and grow them on agar plates as infected colonies (the expected number of plaques is approximately 200//zg of heteroduplex transformed). Hybridize the infected colonies with 32p-labeled deletion primer 1 and plaque purify putative positive clones (optional). 7. Sequence putative positive clones (-20) with an appropriate sequencing primer. 8. Plaque purify and grow up the resultant true positive clone in AC2522 and mutagenize with the second deletion primer and selection primer 2. Prepare overnight cultures of JM 101 for Hanahan transformation the next day. 9. Repeat steps 6 to 8 to isolate and characterize any additional mutants.
Troubleshooting for Ping-Pong Mutagenesis Method Table IV lists problems, possible causes, and remedies useful in PingPong mutagenesis studies. 3I D. Hanahan, J. Mol. Biol. 166, 557 (1983).
[16]
269
M13 VECTORS FOR SITE-DIRECTED MUTAGENESIS
TABLE IV TROUBLESHOOTING FOR PING-PONG MUTAGENESIS METHOD
Problem Low number of transformants (close to background)
Possible causes Polymerase or ligase activity too low
Use flesh polymerase or ligase
Primers degraded
Use fleshly kinased primers Use lower temperature for annealing/extension and ligation reaction Repeat with gel-purified primers Use RNase or PEG ppt to purify DNA templates Redesign primers or use higher temperature for annealing/extension and ligation reaction Remove EcoK or EcoB sites by primers if possible Repeat with plaquepurified template DNA Repeat with cells whose genotype have been reconfirmed Check primers design
Primer too nonspecific
Primers impure High frequency of spurious mutants
All plaques tested were same as parent clone
No plaque observed (even in the control with no primers)
Remedy
Template has too much contaminating RNA Primer has partial homology with several positions along the template Template DNA has EcoK or EcoB sites Template is contaminated with recombinants Incorrect cell strain used
Primer used hybridized to other locations Hanahan procedure failed
Template DNA degraded
Repeat with freshly prepared competent cells Repeat with freshly prepared template DNA
Concluding Remarks F r o m t h e n u m b e r o f p o s i t i v e c l o n e s t h a t h y b r i d i z e d p o s i t i v e l y to the mutagenic oligonucleotide, we have shown that the combined use of the s e l e c t i o n w i t h f o u r c o p i e s o f E c o K / E c o B in t h e v e c t o r / p r i m e r s y s t e m i m p r o v e d the f r e q u e n c y o f h y b r i d i z a t i o n p o s i t i v e c l o n e s a p p r o x i m a t e l y t w o f o l d ( c o m p a r e r o w s 1 a n d 3, 4 a n d 6, a n d 7 a n d 9 o f T a b l e II). T h i s i m p r o v e m e n t is n o t d u e to t h e m e r e p r e s e n c e o f t w o p r i m e r s b e c a u s e t h e u s e o f t h e n o n s e l e c t i v e h o s t TG1 ( E c o K - , E c o B - ) d o e s n o t g i v e a signific a n t i n c r e a s e in t h e f r e q u e n c y o f h y b r i d i z a t i o n - p o s i t i v e c l o n e s , e v e n w i t h
270
MUTAGENESIS AND GENE DISRUPTION
[17]
double primers (compare rows 2 and 3, 5 and 6, and 8 and 9 of Table II). There is no direct relationship between the length of D N A deleted and the percentage o f positive clones obtained (Table II). When the D N A sequences o f the putative positive clones were analyzed, the improvement of the f r e q u e n c y of mutation was even more dramatic (Table III). The percentage of true positive clones as determined by sequence analysis was 30% when selection was used but no mutant was obtained without selection. This result is similar to that obtained when the M I 3 K l l R X vector, which has four tandem copies of EcoK, was used. 4 H o w e v e r , the design of this series of deletion experiments is more flexible than the previous design 4 because the four copies of the EcoK/EcoB site are adjacent to rather than inside the loop of D N A to be deleted. In summary, novel M13 cloning vectors with four overlapping copies o f EcoK (M 13K 119W) and EcoB (M 13B 119) were designed for improving the efficiency o f mutagenesis using oligodeoxynucleotides. These vectors complement the v e c t o r series with four tandem copies of EcoK, M 1 3 K l l R X , which has been shown to be useful in the generation of a series o f unidirectional deletional mutants by a mixture of oligodeoxyribonucleotides. 4,28 Furthermore, we have shown that with the combined use of T4 D N A polymerase and EcoK selection, a high efficiency of mutagenesis of up to 75% can be obtained for large deletions. Acknowledgments I thank the members of the laboratory: V. Li and F. Mui for conducting the sequence analysis, Ken Wong for help in cloning c-fos into the M13 vector. In addition, thanks go to Dr. T. Kunkel for helpful advice and Dr. J. Ferrier for reading the manuscript. K. Wong is a recipient of an MRC summer Farquharson Research Scholarship. This work is supported by a group grant from the Medical Research Council of Canada and a University of Toronto Cannanght research grant.
[17] G e n e S p l i c i n g b y O v e r l a p E x t e n s i o n
By ROBERT M. HORTON, STEFFAN N. H o , JEFFREY K. PULLEN, HENRY D. HUNT, ZELING CAI, and LARRY R. PEASE Introduction Conventional methods o f engineering recombinant D N A make use o f restriction enzymes to cut molecules apart at specific nucleotide METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
[17]
GENE SPLICING BY OVERLAP EXTENSION
271
sequences and ligases to rejoin the parts. A significant limitation of this technology is that restriction enzymes are sequence dependent and these recognition sequences appear more or less randomly in DNA. That is, restriction enzymes cut where recognition sites are located and not necessarily at optimal positions along the gene for purposes of genetic engineering. The polymerase chain reaction (PCR) has made possible a sequence-independent engineering method that we have referred to as "gene splicing by overlap extension" or " S O E . " This technology is especially useful in complicated constructions that require precise recombination points, such as joining two coding sequences in frame, and it also provides a straightforward way of performing site-directed mutagenesis. ~,2
Method The basic scheme of gene splicing by overlap extension is illustrated in Fig. 1. 3 The process requires two steps. First, the specific fragments to be joined are isolated by PCR. The ends of the amplified fragments are modified during this step so that the two fragments "overlap," or share complementary sequences on the strands to be joined. Following denaturation and reannealing, strands from the two fragments act as primers on each other. Extension of the overlap by DNA polymerase results in the recombinant product. A detailed depiction of the overlap region is shown in Fig. 2). The first step in the SOE reaction is an application of "mispriming," in which extra, unrelated sequences added to the 5' end of a PCR primer become incorporated into the end of the product. 4 It is a conventional reaction, but it uses specially designed primers. The second step (overlap extension) is as simple to carry out as the first; it just requires the two purified fragments to be put together under"PCR conditions," with buffer, dNTPs, polymerase, and thermal cycling. Only one strand from each of the original PCR products is actually incorporated into the final product. The two strands act as primers on each other to form a single fused molecule. Inclusion of PCR primers for the distal ends of each fragment allows the final product to be amplified.
1 R. M. Horton, H. D. Hunt, S. N. Ho, J. K. Pullen, and L. R. Pease, Gene 77, 61 (1989). 2 S. N. Ho, H. D. Hunt, R. M. Horton, J. K. Pullen, and L, R. Pease, Gene 77, 51 (1989). 3 R. M. Horton, Z. Cai, S. N. Ho, and L. R. Pease, BioTechniques 8, 528 (1990). 4 K. Mullis and F. Faloona, this series, Vol. 155, p. 335.
272
MUTAGENESIS AND GENE DISRUPTION
[17]
G e n e II
PCR 2 J,~l II!11~1 U
_ _ _ J
. . . .
AP~
/
HecomDlnant proouct FIG. 1. The general concept of overlap extension. First, two PCR products are made in separate reactions; primers a and b produce product AB from gene I and primers c and d are used to amplify fragment CD from gene H. DNA segments are depicted as paired antiparallel strands. An arrowhead indicates the 5'-to-3' direction of each strand of the primers and PCR products; the ends of the template genes are not shown. Primers b and c have had sequences added to their 5' ends so that the right end of AB matches the sequence at the left end of CD. When these products are mixed in an SOE reaction, the top strand of AB "overlaps" with the bottom strand of CD, their 3' ends being oriented toward each other. This allows them to act as primers on one another to make a giant "primer dimer," which is the recombinant product. The other strands, which point in the wrong directions, do not form product and are not necessary to the reaction. (Reprinted from Horton e t al. 3 with permission from the publisher.)
GENE SPLICING BY OVERLAP EXTENSION
[17]
273
Product AD (433 bp)
~
AGAGGTCAAATTCCACC-~ TCTCCAGTTTAAGGTGG"s' +
~"AGAGGTCAAATTCCACC~ , ~
S'-TCT COAGTT T AAGGT(~:~~ , ~ " ~",,~" Product EH (581 lop)
........................
.~.:..~...~
'~
................... i
~'~ i ~
~
/~AGAGGTCAAATTCCACC.-~~ ~r'TCTCCAGTTTAAGGTO~
~'%
e
i
i ........................
~
..................
AGAGGTCAAATTCCACC~--~ J P ' ~ ~ TCTCCAGTTTAAGGT~,~ ~ 7 Reeomi~antProduct
"
X
(997 bp)
FIG. 2. Detail of ends of fragments being spliced together. In this figure, the two products being joined are AD and EH. The design of this construct is discussed elsewhere) Strands with the 5'-to-3' direction going left to right are shown in white, and the opposite strands are shown in black. The right end of AD has the same sequence as the left end of EH. When AD and EH are mixed in an SOE reaction, heated, and reannealed, the 3' end of the white strand of AD overlaps with the 3' end of the black strand from EH, Extension of this overlap by DNA polymerase creates the recombinant product.
Examples Site-Directed Mutagenesis A simple example of the use of overlap extension is for site-directed mutagenesis. As illustrated in Fig. 3, the same template is used to make both fragments AB and CD. Differences from the template sequence are introduced in the primers that generate the overlap regions in the amplified fragments generated in the primary PCR (Fig. 4A).
Splicing Genes Together An example of gene splicing by overlap extension is the construction of a chimeric class I major histocompatibility complex molecule using parts from two different members of this multigene family. The purpose was to examine the functional role of the a helices in these molecules by switching them from one molecule to the other. This construction is complex because it involves splicing together four fragments in consecutive reactions, as illustrated in Fig. 5. Notice also that the segments originated from different exons in the original templates, but that the intron has been
274
MUTAGENESIS AND GENE DISRUPTION
[17]
C
(1) ~ a~
b
(2) 1 c+~
d
AB CO (a)
~ ~ NB4.CO
. . . . .
oo...
o . . . . . e o ~ .
J
MUTANT FU~ON PRODUCT
FIG. 3. Mutagenesis by overlap extension. In simple mutagenesis, both of the products to be joined are amplified from the same template. Changes to the original sequence, represented by the black rectangle, are made by "mispriming" using primers b and c. (Reprinted from Ho e t al. 2 with permission from the publisher.)
A
5' CGGTACATGTCTGTCGGCTACGTC 5' .... GCCCCGGTACATGGAAGTCGGCTACGTC.... .... CGGGGCCATGTACCTTCAGCCGATGCAG.... 5' CGGGGCCATGTACAGACAGCCG 5'
B
template
5' TCCCTGCGGCGGCTGCGCACAGGTGC 3' 5' AGGGACGCCGCCGA 5'
FIG. 4. Examples of primers used in overlap extension reactions. (A) Site-directed mutagenesis: The relationships between the mutagenic oligonucleotides and the template are indicated. Mutations to be introduced are indicated by asterisks. (B) Gene splicing by overlap extension: The regions of the oligonucleotides that allow them to act as primers on the appropriate template are underlined. The complementary regions of overlap between the pairs of primers that allow them to be spliced together are indicated by asterisks.
[17]
GENE SPLICING BY OVERLAP EXTENSION
275
deleted in the final construct. This precise "splicing" of exon sequences is a trivial exercise using SOE, but the reader is invited to consider how difficult such a construction might be using conventional restriction enzyme methods, unless useful sites are serendipitously present at the exon boundaries. General Considerations In designing a project using SOE, it is essential to keep the strands and their orientations straight. The notation used in Figs. 1, 3, and 5 is useful in this regard. Each strand is represented by an arrow that indicates the 5'-to-3' direction. This is the direction in which DNA polymerase can extend the strand when it acts as a primer. By convention, DNA sequences are usually written out so that the reading frame of reference reads left to right, and the top strand has its 5' end at the left and its 3' end at the right. If only one strand is shown, it is assumed to be the top strand. When designing SOE primers, it is useful to write out the sequence of both strands of the recombinant product (or mutant) wanted, as in Fig. 4, and to copy the primer sequence from this. It is important to design the primers carefully because their sequences must be exactly correct. Although these words of caution may sound superfluous, it has been our experience that most errors associated with this approach occur in the planning phase. Each primer extension event requires sequence homology between the primer and the template to permit hybridization. We have empirically designed our primers to overlap with the template by approximately 16 to 20 bp. In our experience, oligonucleotides with a minimum 16-bp region of homology have consistently provided adequate amplifications. Given the decreasing cost of oligonucleotide synthesis, we have not spent significant effort in working out parameters such as the minimum number of overlapping nucleotides or the base composition of the overlapping region that is required for SOE or mutagenesis using this approach. In general, we try to avoid sequences with significant internal homology. When given the option, we choose sequences containing more or less equal amounts of all four nucleotides. An example of an overlapping region we have used successfully is shown in Fig. 4B. In the few cases in which a given oligonucleotide did not provide adequate amplification, resynthesis of the same sequence usually resulted in successful amplifications. Gene splicing by overlap extension is clearly the method of choice for certain constructs that would be difficult to carry out using conventional techniques such as restriction enzyme digestion and ligation. A common example of such a project is engineering fusion proteins, in which the reading frame and the sequence of the desired protein place rigid con-
MUTAGENESIS AND GENE
276
[17]
DISRUPTION
A exon 2
~,y
infron 2
exon 3
y....~
PCRI (1) fra~,n~ AS
.
~"I~
ePCR~(3)~-'"f fraQment ~..
..
_
c( heix
~
¢*----1""~
c( heix v . H_2Ld 81~w~lfleetrWnolleklB
t -,
'A
•
~ X
H-2K 10......... t"............... r.r........-~......... ..w.......~...w.L...*............................. ..~m.......w............ .~.....~....r.~............. .~....:~
c F'CR~(2)'"-d
g
PCR~(4)
T
h
T
. . . . . . . . . . . . . . . . . . . . . . . . . .--% ~.__
~=====================:-=============l====== ~.
fragment CO
fragment GH
B fragnmnt AB
T" . . . . . . . . . . . . . . . . . . . . . . .
fragment CD
SOEI( 1) fragm6nt EF b/c joint ..................... ~ fragment AD .::. i SOEI(3) exon 2 Z 09
a-.-~
:'::";'":;':;;;;;":;"::";;':;':;":':: I fragment GH SOE~(2) f/g joint =='=;=';;=".==='"';=;'=;='=;'"==;='='='; fragment EH exon 3
oc helix Y:mtron deleted at die joint
:::~:::::,*::7,:',::::
o¢helix ~ X
:::::::::::::::::::::::::::::::::::
e.
.e.~
g..
......
~":
':'-
~":
"' h
b
"d f Recombinant Molecule
FIG. 5. Strategy for c o n s t r u c t i n g a c o m p l e x fusion protein by splicing parts together. (A) P C R - g e n e r a t e d f r a g m e n t s . (B) S O E reactions. (Reprinted f r o m H o r t o n e t al. i with p e r m i s s i o n f r o m the publisher.)
straints on how the recombination can be made. However, SOE has certain drawbacks that make it less appealing for "ordinary" applications. First, because anything done with PCR involves in vitro synthesis of DNA (both chemical and enzymatic), there is an increased probability of errors being introduced. Thus, for many applications, especially when the product is
[17]
GENE SPLICING BY OVERLAP EXTENSION
277
to be cloned, the possibility exists that several clones may have to be sequenced to find one that is entirely correct. We have found error frequencies (i.e., the proportion of incorrect nucleotides in the final product) between 0.026 and 0.06% in overlap extension reactions using Taq DNA polymerase. ~,2 Other polymerases with lower error rates would presumably lower this frequency even further. The cloning strategy used in the examples shown here is a "cassette" approach, in which the final product is cloned into a plasmid vector in the conventional manner. Several groups have demonstrated that products can be spliced directly into vectors. 5-8 While these techniques remove some of the limitations inherent in using restriction enzymes, and may simplify and speed up the cloning process, the cassette approach has the advantage that only the cassette portion of the final construct has been subjected to DNA synthesis in vitro, limiting the amount of sequencing that must be done to be sure of having an error-free clone. Other interesting examples of this and related technologies have appeared in the literature. 9-12
Protocols
Solutions and Reagents Standard PCR buffers and conditions are suitable for SOE. Because a high [Mg z+ ] appears to lead to increased rates of misincorporation by Taq polymerase,13 it is advisable to use the lowest concentration of magnesium compatible with amplification of the specific segments of interest. Generally, a titration from 0.5 to 2.5 mM will reveal a range of [Mg 2+] that gives good amplification; working on the low end of this range should result in lower error rates. For these titrations, it is convenient to have the MgC% separate from the buffer [10 x buffer is 500 mM KC1, 100 mM Tris-HC1 (pH 8.3)]. 5 D. H. Jones and B. H. Howard, BioTechniques 8, 178 (1990). 6 A. R. Schuldiner, K. Tanner, L. A. Scott, C. A. Moore, and J. Roth, Anal. Biochem. 194, 9 (1991). 7 G. S. Sandhu and B. K. Kline, Minn. PCR Symp. abstract and poster (1991). 8 A. R. Schuldiner, L. A. Scott, and J. Roth, Nucleic Acids Res. 18, 1920 (1990). 9 G. Sarkar and S. S. Sommer, BioTechniques 8, 404 (1990). ~0R. M. Horton and L. R. Pease, in "Directed Mutagenesis: A Practical Approach" (M. J. McPherson, ed.), p. 217. IRL Press, Oxford, England, 1991. u B. Berkhout, A. Gatignol, A. B. Rabson, and K.-T. Jeang, Cell 62, 757 (1990). 12 C. Abate L. Patel, F. J. Rauscher III, and T. Curran, Science 249, 1157 (1990). 13 K. A. Eckert and T. A. Kunkel, Nucleic Acids Res. 18, 3739 (1990).
278
MUTAGENESIS AND GENE DISRUPTION
[17]
Polymerase Chain Reaction Conditions Polypropylene tubes (0.6 ml) (Robbins Scientific, Sunnyvale, CA) Native Taq DNA polymerase (Perkin-Elmer-Cetus, Norwalk, CT) Ultrapure dNTP mix (Pharmacia, Piscataway, N J) Mineral oil, light (Sigma, St. Louis, MO) DNA thermal cycler (we use a model from Perkin-Elmer Cetus); for 25 cycles: Denaturation at 94 ° for 1 min Annealing at 50 ° for 2 min Extension at 72 ° (1 to 3 min, as determined by an enzyme rate of 1000 bases/min) The shortest denaturation and annealing temperatures required will vary from instrument to instrument. Very short cycling times have been used successfully by some investigators 14
Overlap Extension 1. Amplify the intermediate products AB and CD in separate tubes: Template, 100-500 ng/100/zl Primer 1, 1/zM Primer 2, 1 IzM Buffer, 1 x Mg 2+, empirically determined dNTPs, 200/~M Taq polymerase, 0.025 U//xl Final volume, 100 tzl Mineral oil: Add 2-3 drops to assembled reaction mixtures In the first reaction, the template is gene I and the primers are a and b. This produces product AB. In the second reaction, to produce product CD, the primers are c and d, and the template is gene H. For simple sitedirected mutagenesis, both reactions use the same template. A broad range of template concentrations will work. On theoretical grounds, we recommend starting with as much as 0.5/~g of template, if possible. This reduces the number of rounds of replication required to amplify a workable amount of product. Too much template DNA can inhibit the reaction, however; for plasmids we have found that concentrations above 1 tzg/100 ~1 commonly cause inhibition. The actual number of rounds of synthesis (doublings) is not the same as the number of cycles of heating/cooling the sample has been exposed to; once the reaction reaches 14 C. T. Wittwer, G. C. Fillmore, and D. J. Garling, Anal. Biochern. 186, 328 (1990).
[17]
GENE SPLICING BY OVERLAP EXTENSION
279
a maximum, further cycling has little effect. Therefore, we have made no attempt to minimize the number of cycles, and generally let it go for 20 or 25. 2. Gel-purify products AB and CD: Fragments larger than about 300 bp can be electrophoresed through an agarose gel in TAE buffer [40 mM Tris-acetate, 1 mM ethylenediaminetetraacetic acid (EDTA), pH 8.2], then the desired band can be cut out and the DNA recovered by the glass bead elution procedure (GeneClean; Bio 101, La Jolla, CA). For small fragments, better resolution can be obtained with high-percentage agarose gels [1% (w/v) normal agarose plus 1-3% NuSieve GTG agarose (FMC BioProducts, Rockland, ME)]. Fragments smaller than about 200 bp cannot be recovered efficiently from glass beads, so an alternative procedure should be used, such as running the band into a well cut in the gel, 15adding 10/zg yeast tRNA (Bethesda Research Laboratories, Gaithersburg, MD) as a carrier, and precipitating with ethanol. 15 Other workers use less extensive purification schemes, ~6 or do not purify the intermediates at all. ~7 In our experience, gel purification of the intermediates reduces the background of unwanted side products, sometimes quite dramatically. 3. SOE fragments AB and CD: The SOE reaction is done under the same conditions as the PCRs, except that two templates are used instead of one: Template 1 (product AB), - 2 5 % of total Template 2 (product CD), - 2 5 % of total Primer 1 (primer a), 1 ~M Primer 2 (d), I /xM Buffer, 1 × Mg 2+, emperically determined dNTPs, 200/xM Taq polymerase, 0.025 U//xl Again, a wide range of template concentrations will work, but larger amounts of template theoretically will lead to lower error frequencies. We recommend using about one-fourth of the purified product from a 100-/sl reaction to permit additional reactions, should the first be unsuccessful. Although we usually use roughly equimolar amounts of the two SOE templates, this is not necessary. The reaction will work even if one template is in gross excess over the other. ~5T. Maniatis, E. F. Fritsch, and J. Sambrook, "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982. 16 R. Higuchi, B. Krummel, and R. K. Saiki, Nucleic Acids Res. 15, 7351 (1988). 17 j. Yon and M. Fried, Nucleic Acids Res. 17, 4895 (1989).
280
MUTAGENESIS AND GENE DISRUPTION
[18]
[18] S e l e c t i o n o f O l i g o d e o x y n u c l e o t i d e - D i r e c t e d M u t a n t s By CARL A. BATT, YUNJE CHO, and ANDREW C. JAMIESON Introduction Mutational analysis has been used for many years to probe the genetic basis of a given phenotype. Prior to advances in in vitro nucleic acid enzymology this was usually accomplished by treating the targeted organism with a chemical or physical agent to enhance the rate of mutation. Depending on the particular mutagenic agent, base substitutions, deletions, or insertions could be anticipated. The challenge was then to devise a suitable screening protocol to identify mutations of interest and characterize them within the limits of the available biochemical methods. The ability to manipulate nucleic acids in vitro, coupled with techniques for precisely determining changes to a given sequence, has permitted the directed mutation of DNA. Although methodologies for chemical and random enzymatic mutagenesis have been established, where the target sequence and desired change can be defined, oligonucleotide site-directed mutagenesis is usually the method of choice. Examples of the potential of site-directed mutagenesis include probing the role of specific amino acid residues in the structure of a protein, incorporating desired restriction sites, and studying the effect of altering a nucleotide sequence on binding of a regulatory protein.
Principle of Method Oligonucleotide-mediated site-directed mutagenesis allows the selective substitution, insertion, or deletion of one or more targeted nucleotides. 1 A mutagenic primer containing the desired nucleotide sequence change is annealed to a template and the remainder of the sequence is synthesized enzymatically. For single-stranded templates (i.e., M 13mp 19) one would intuitively assume that the population of molecules recovered by transformation would be a ratio of 50 : 50, wild type : mutant. This in practice is never the case and the frequency of mutants can be as low as <0.0001%. There are a number of factors contributing to this low frequency of mutant recovery, including failure to complete second-strand I D . B o t s t e i n a n d D . S h o r t l e , Science 2 2 9 , 1193 (1985).
METHODS IN ENZYMOLOGY, VOL. 217
Copyright © 1993 by Academic Press, Inc. All rights of reproduction in any form reserved.
[18]
OLIGONUCLEOTIDE-DIRECTED MUTANTS
281
synthesis and perhaps a biological bias against the in vitro synthesized mutant strand. Several different approaches have been developed for selective enrichment of the mutant strand either in vivo or in vitro. The basic strategy is to introduce a chemical modification that distinguishes the parental (nonmutant) strand from the mutant strand. This modification is coupled to a biological selection either in vitro or in vivo for the mutant strand and consequently to eliminate the parental strand. In one case, a mutant host strand of Escherichia coli is used to produce a parental template that is selected against during transfection into nonmutant E. coli strains. 2 The second approach to enrich for mutations involves the synthesis of the mutant strand using deoxynucleotide analogs that ensures that the mutant (but not the parental) strand is resistant to certain restriction enzymes. The parental strand is nicked and removed, using an exonuclease. An example is the incorporation of phosphorothioate into the mutant strand, which renders it resistant to cleavage by N c i I . 3'4 The protocol developed in our laboratory involves synthesizing a mutant strand that is resistant to restriction? The second strand is synthesized from a mutagenic primer using the standard deoxynucleotides, except 5-methyldeoxycytodine triphosphate (d5'-MeCTP) replaces dCTP. The d5'-MeCTP renders the mutant strand resistant to a number of restriction enzymes, including MspI and Sau3A. These enzymes share the ability to nick the nonmethylated strand of a hemimethylated DNA molecule.6 The enzyme HhaI is added to this restriction mixture because it can cleave any single-stranded (uncopied) parental template. The nicked parental strand is then removed using exonuclease III. There is no need to resynthesize the exonuclease-digested strand and regenerate a double-stranded molecule before transformation. The methylated single-stranded mutant DNA is then transformed into an E. coli m c r A B - strain. The m c r A B mutant does not restrict the d5'-MeCTP DNA as compared to other strains of E. coli. The resulting plaques (or colonies when phagemids or plasmids are used) are screened for the desired mutation by sequencing using a second primer that binds approximately 50-75 bp from the site of the mutation. Alternatively, where a restriction site is to be introduced, the replicative form is isolated and subjected to restriction analysis. 2 T. A. Kunkel, J. D. Roberts, and A. Zakour, this series, Vol. 154, p. 367. 3 K. L. Nakamaye and F. Eckstein, Nucleic Acids Res. 14, 9679 (1986). 4 j. W. Taylor, W. Schmidt, R. Cosstick, A. Okruszek, and F. Eckstein, Nucleic Acids Res. 13, 8749 (1985). 5 M. A. Vandeyar, M. P. Weiner, C. J. Hutton, and C. A. Batt, Gene 65, 129 (1988). 6 M. Nelson and M. McClelland, Nucleic Acids Res. 15 (Suppl.), r219 (1987).
282
MUTAGENESIS AND GENE DISRUPTION
[18]
Materials and Reagents
Bacterial Strains. Any one of a number ofE. coli strains that carry the mcrAB- mutation can be used, including JM2r-,5 SDM (U.S. Biochemical, Cleveland, OH), SURE (Stratagene, La Jolla, CA), DH5aMCR (BRLLife Technologies, Gaithersburg, MD), and ER1451, (available from New England BioLabs, Beverly, MA). For all strains that do not carry an F' necessary for replication of M13, an indicator strain must be used but this strain need not be a mcrAB- derivative. Primers. The synthetic oligonucleotide primers are designed to have 8-10 bp flanking the mismatched region. Routinely the primers are used directly after deprotection and do not need any further high-performance liquid chromatography (HPLC) purification. Enzymes and Reagents. The initial methodology we developed used T4 DNA polymerase (U.S. Biochemical or Pharmacia, Piscataway, N J). Subsequently, T7 DNA polymerase (native) was tested and found to be superior with respect to its rate of second-strand synthesis. Although a number of restriction enzymes will cleave the nonmethylated strand of hemimethylated DNA we routinely use MspI [Boehringer Mannheim (Indianapolis, IN), New England BioLabs, U.S. Biochemical] and IthaI to cleave any residual single-stranded template. Exonuclease III is purchased from U.S. Biochemical and d5'-MeCTP is obtained from Pharmacia.
Method The following is a detailed protocol for the individual steps in our sitedirected mutagenesis system. Each step can be carried out sequentially without any precipitation or phenol extraction between steps. The enzymes, where necessary, are heat inactivated between each step. Kinasing of Mutagenic Oligonucleotide Primer. The first step is to kinase the primer in order to permit the enzymatic extension of the nucleotide sequence. It is possible to have this step performed chemically when the primer is synthesized, obviating the need for an enzymatic phosphorylation. The reaction is carried out with 50 pmol of primer, 100 mM TrisHCI (pH 8.0), l0 mM MgCI2, ! mM ATP, using 1 unit T4 kinase in a total volume of 50/zl. The reaction is carried out at 37° for 30 rain and then the enzyme is inactivated at 70 ° for l0 min. Annealing Primer to Template. The kinased primer is then annealed to the single-stranded target. Approximately 3 pmol of kinased primer is added to 2 /zg of template DNA in 20 mM Tris-HC1 (pH 7.5), 10 mM MgCl~, 50 mM NaC1 in a total volume of 20/zl. The mixture is heated to 65 ° for 5 min and then cooled slowly to room temperature.
[18]
OLIGONUCLEOTIDE-DIRECTED MUTANTS
283
Second-Stand Synthesis. The second strand is synthesized enzymatically, during which d5'-MeCTP is incorporated. To the annealed product dATP, dGTP, dTTP, and d5'-MeCTP (prepared in Tris-HCl, pH 7.5) at a final concentration of 0.5 mM, 1 mM ATP, 2 units T4 ligase, and 5 units of T4 (or T7) DNA polymerase are added in a final volume of 50/zl. The reaction is incubated at 37° for 90 min, then at 70° for 10 min to inactivate the polymerase and ligase. Removal of Parental Strand by Restriction and Nuclease Digestion. The parental strand is removed on the basis of the ability of MspI to restrict this nonmethylated strand. To the heat-inactivated second-strand synthesis mixture I0 units HhaI and 10 units Msp! are added and the reaction incubated at 37° for 1 hr. Next, 100 units ExoIII is added and the mixture is incubated at 37° for 45 min, then inactivated at 70° for 10 min. The final reaction mixture is approximately 55/~1, depending on the unit activity of the enzymes used. Transformation. Approximately 2 ~1 of the ExoIII reaction is used without any further treatment to transform E. coli JM2r- as per standard M13 transfection protocol using calcium chloride-treated cells. 7 Competent cells can also be generated by electroporation 8 with efficiencies on the order of 103 times greater than for calcium chloride treatment. Briefly, 500 ml of L broth is inoculated and grown to a n OD600 of approximately 0.5. The cells are washed with sterile distilled water and concentrated 500fold. These cells can be stored at - 70° in 10% (v/v) glycerol and used for electroporation at 12.5 kV/cm, with a 25-/~F capacitor and a 200 f~ (ohm) in series. The electroporated cells are then incubated for 30 min at room temperature and plated in a top agar layer with an appropriate indicator strain. Again it is not necessary that the F + indicator strain be an mcrAB strain. Results. The site-direct mutagenesis protocol described here has been used to produce a considerable number of mutations containing up to four consecutive base-pair mismatches. Mutations have been recovered at frequencies ranging from 14 to 100%. Table I is a listing of the mutagenic primers and the frequency at which those mutations were recovered that has been compiled from our own laboratory in addition to others. In certain cases in which primers with wildcard codons (all 64 possible codons at the target incorporated at random during primer synthesis) have been used, not all of the expected mutations are recovered. This occurred, for example, with primer 7 T. Maniatis, E. F. Fritsch, and J. Sambrook, "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1982. 8 W. J. Dower, J. Miller, and C. W. Ragsdale, Nucleic Acids Res. 16, 6127 (1988).
284
[18]
MUTAGENESIS AND GENE DISRUPTION TABLE
I
MUTAGENIC PRIMERS, SEQUENCES, AND FREQUENCIES Primer USB 1 USB2 J 13 J 16 B L G C 121S BLGV104C B L G A 132C BLGF136A XIW49A XlW188F XlW188A XIH101Q X I H 101S XIH101N XIH271N XIT141A
Primer sequence a GGTTTTCCTTAGTCACGA GCCCAGGGTTTTCCCATGGGTCACGACGTTGTAAAA AAGCAAGTTGTGGACG GGAAGCAAGTTGTGGACGTC TGCCAGTCCCTGGTCAG AAGTACCTGT_...GCTTCTGCATG GACGAGT__.GGCCTGGAGA TGGAGAAAG___C_CCGACAAAG GGTGTCGCGTAGCAG GACCGCCTAQCAGGAC ACCGCCCGCCACGAC CCACATCTTGGAAGCAAT
CCACATCGGAGAAGCAAT ATCCACATCNNNGAAGCAATA GCGTCGCNNNGTTAGCTTC GTTGGCGCGTCCCCA
AG b -
30.4 79.7 29.3 36.8 36.2 35.9 31.0 28.6 22.7 36.6 37.4 36.6 32.6 37.6 33.0 38.9
Percentage e 79 50 50 50 d 100 14 66 66 14 e 100 80 75 40 90 f 80 g 20
o Sequence of mutagenic primer. Nucleotide(s) underlined are mismatches to targeted sequence. b kcal -+ 10%. F r o m I. T i n o c o , P. N . B o r e r , B. D. D e n g l e r , M . D . L e v i n e , O . C. Uhlenbeck, D . M . Crothers, and J. Gralla, Nature (London) New Biol. 246, 40 (1973). c Frequency of desired mutation as determined by sequence analysis. d Single-base pair deletion of mismatched T observed in 17% of the plaques. e Single-base pair deletion of mismatched C observed in 33% of the plaques. f O n l y two mutant codons, TAT and CGT, were recovered. g A total of 14 different mutant codons was recovered.
XIH101N (Table I), where only TAT and CGT, in addition to the nonmutant codon CAC, were obtained. An examination of the potential hairpin structures within the mutagenic primer revealed that TAT and CGT disrupted a potential hairpin, therefore allowing those primers to bind more efficiently. In contrast, a total of 14 different codons was obtained using the XIH271N primer. Another observation was the occurrence of single-base pair deletions (primer XIW188F and J16). In those cases the deletions occurred within the mismatch and fell at the apex of a small hairpin. In experiments in which these deletions occurred, the desired mutation was also recovered and the deletion was observed approximately 33% of the time. Predicting the appearance of these deletions is difficult because in the case of primers J13 and J16 the deletion was observed only in J16, whose sequence includes primer J13, for which no deletions were observed.
[18]
OLIGONUCLEOTIDE-DIRECTED MUTANTS
285
Conclusions
The methodology described above has been in place in our laboratory for the past 4 years. In addition to being used for substituting nucleotides, it has also proved efficient for inserting and deleting specific nucleotides (P. Flicke, personal communication, 1992). Most of the problems that we have encountered relate to the transformation frequency of the E. coli JM2r strain using calcium chloride. Therefore, we usually employ electroporation to transform this strain. (Frozen competent E. coli SDM cells are, however, available from United States Biochemical.) The selection of DNA polymerase is predicated on the absence of any strand displacement activity, which may result in the dislodging of the mutagenic primer after the first round of second-strand synthesis, leading to the synthesis of a subsequent second-strand molecule without the appropriate mutation. Although T4 DNA polymerase has proved satisfactory, efforts have indicated that the number of progeny (but not the frequency of mutations) is increased when T7 DNA polymerase is used (J. Webb, personal communication, 1990). 9 This is presumably due to the ability of T7 DNA polymerase to be highly processive, incorporating thousands of nucleotides before dissociating from the template. This ensures that the majority of the template molecules are fully copied.l° We have not exhaustively compared the performance of the enzymes (polymerases, restriction enzymes, and exonucleases) from different suppliers. The choices listed under Materials and Reagents (above) are only suggestions. Because the MspI and HhaI restrictions are carried out in the same buffer, some problems may be encountered with any given enzyme from any given supplier. We have found that the enzymes available from the suppliers cited above are compatible but that does not preclude the potential utility of enzymes from other sources. In our laboratory we have successfully obtained all site-specific substitutions attempted, although this does not imply that the methodology is infallible. Clearly, the annealing conditions described are not optimized (nor in fact modified) for each individual primer sequence. In retrospect for the case of XIH101N, a wider distribution of substitutions might have been obtained using either higher annealing temperatures or including a small concentration of formamide to reduce secondary structure. There is no obvious correlation between the AG for a given primer and the mutation frequency (Table I). In fact, for the only easily comparable case, J13 vs 9 C. A. Batt, J. Webb, P. Oren, and P. Flick, BioTechniques 9, 555 (1990). ~0 R. Lechner, M. Engler, and C. C. Richardson, J. Biol. Chem. 258, 11174 (1983).
286
MUTAGENESIS AND GENE DISRUPTION
[19]
J16, although the latter has a lower AG by 7.5 kcal, there was no difference in the frequency at which mutations were obtained. The methodology for carrying out site-directed mutagenesis has evolved greatly over the past few years. Beyond techniques based on similar methodology as described here, the most significant advance has been in the application of the polymerase chain reaction (PCR) to generate site-directed mutations. A number of protocols have been developed using either two or three oligonucleotide primers. Due to the practical limits of the PCR using Taq DNA polymerase, only a short region is amplified and then cloned into a recipient vector. One potential problem, which has not been extensively documented, is the known lack of fidelity of these thermostable DNA polymerases, especially Taq DNA polymerase, which may introduce second-site mutations in the target. Acknowledgments The authors acknowledge the contributions of Clara Finch, Jonathan Miller, Parke Flick, and Jennifer Webb for supplying data on the mutation frequencies. This work was supported by a grant from the Cornell Biotechnology Program, which is sponsored by the New York State Science and Technology Foundation, a consortium of industries, the U.S. Army Research Office, and the National Science Foundation.
[19] C h e m i c a l C l e a v a g e o f M i s m a t c h to D e t e c t M u t a t i o n s By JENNIFER A. SALEEBA and RICHARD G. H. COTTON Introduction The chemical cleavage of mismatch (CCM) method allows mutation sites to be detected in kilobase-length pieces of nucleic acids. J A screening method such as this obviates the need to sequence lengths of DNA to determine mutation sites. The method can be widely applied. It can be used to compare a clone of known or unknown sequence to samples of mutant origin, and for applications such as confirmation of in vitro mutagenesis. It has been
J R. G. H. Cotton, N. R. Rodrigues, and D. R. Campbell, Proc. Natl. Acad. Sci. U.S.A. 85, 4397 (1988).
METHODS IN ENZYMOLOGY.VOL. 217
Copyright© 1993by AcademicPress, Inc. All rightsof reproductionin any formreserved.
[19]
CHEMICAL CLEAVAGE OF MISMATCH METHOD
287
widely applied to diagnosis of human inherited disease 2-7 and to mutation detection in other organisms. 8-1°
Principle of the Chemical Cleavage Method Chemical cleavage is based on the principle that mismatched or unmatched residues are more reactive to modification by the chemicals hydroxylamine and osmium tetroxide than matched bases. This can be used to advantage for detection of point mutations and small insertions and deletions. Heteroduplex molecules can be formed by annealing a control and a test piece of DNA. Sites at which residues are mismatched or unmatched will be modified by the chemical hydroxylamine, in the case of cytosine residues, and osmium tetroxide, in the case of thymine residues. Cleavage of the DNA at the site of the modified residues is achieved by subsequent reaction with piperidine. Cleavage products are resolved by denaturing gel electrophoresis, and detection usually takes place by autoradiography (Fig. I). If necessary the exact change present in a mutant is then determined by sequencing the region identified by CCM. Mismatched or unmatched guanine and adenine residues are detected by virtue of the fact that their complementary cytosine and thymine residues are reactive. Cytosine and thymine residues surrounding mismatch sites may have a limited reactivity with hydroxylamine and osmium tetroxide, due to disruption of the helix.11 Analysis of a number of DNA fragments screened for mutations by chemical cleavage has shown that some T residues in T. G mismatches are resistant to modification by osmium tetroxide, and subsequent piperidine 2 j. F. Bateman, S. R. Lamande, H.-H. M., Dahl, D. Chan, T. Mascara, and W. G. Cole, J. Biol. Chem. 264, 10960 (1989). 3 M. Grompe, C. T. Muzny, and C. T. Caskey, Proc. Natl. Acad. Sci. U.S.A. 86, 5888 (1989). 4 A. J. Montandon, P. M. Green, F. Giannelli, and D. R. Bentley, Nucleic Acids Res, 17, 3347 (1989). 5 D. W. Howells, S. M. Forrest, H.-H. M. Dahl, and R. G. H. Cotton, Am. J. Hum. Genet. 47, 279 (1990). I. Dianzani, S, M. Forrest, C. Camaschella, G. Saglio, A. Ponzone, and R. G. H. Cotton, Am. J. Hum. Genet. 48, 631 (1991). 7 S. M. Forrest, H.-H. M. Dahl, D. W. Howells, I. Dianzani, and R. G. H. Cotton, Am. J. Hum. Genet. 49, 175 (1991). 8 R. G. H. Cotton and P. J. Wright, J. Virol. Methods 26, 67 (1989). 9 M. Han and P. W. Steinberg, Cell 63, 921 (1990). l0 M. Grompe, J. Versalovic, T. Koeuth, and J. R. Lupski, J. Bacteriol. 173, 1268 (1991). tl R. G. H. Cotton and R. D. Campbell, Nucleic Acids Res. 17, 4223 (1989).
288
MUTAGENESIS AND GENE DISRUPTION
--T --A
G C
--G ----C
[19]
C G,
I formation Heteroduplex -~C
#
G
~A
C#--'~
I Cleavage H
OT
FIG. 1. Principle behind chemical cleavage. Wild-type and mutant DNAs are heated and annealed to give heteroduplex molecules. Mismatched T resides are reactive to osmium tetroxide (*) and mismatched C residues are reactive to hydroxylamine (#). Cleavage of DNA strands occurs at these sites by reaction with piperidine. Cleavage products are resolved by denaturing gel electrophoresis, and detected by autoradiography or silver staining. A probe band in the first lane and cleavage products after reaction with hydroxylamine (H) and osmium tetroxide (0T) are indicated.
cleavage. 12 The sequence context surrounding the mismatch appears to be important in determining whether a particular T. G mismatch will be modified by osmium tetroxide. Fortunately, a heteroduplex containing a T. G mismatch is always accompanied by a complementary heteroduplex containing an A. C mismatch. An unreactive T- G mismatch can therefore be easily detected by the complementary A. C mismatch, which is highly reactive to hydroxylamine. With this approach, a strategy offering complete detection of mutations by chemical cleavage can be achieved. When using radiolabeled detection, probes are made from both test and control DNAs and used either separately or together in CCM reactions. If using 12 R. G. H. Cotton, manuscript in preparation.
[19]
CHEMICAL CLEAVAGE OF M1SMATCH METHOD
289
a nonradioactive mode, where all DNA is detected, one set of CCM reactions will allow all mutations to be identified. Internally and end-radiolabeled probes and unlabeled DNA may be used to detect cleavage products in CCM. End-labeled probes will be cleaved to reveal one detectable cleavage product for each mutation. This type of probe is best applied to examples in which a number of mutations are expected. If the exact position of the mutation must be known, a single strand should be labeled. Internally radiolabeled probes will give two cleavage products per mutation. These probes are easy to prepare and are therefore well suited to mutation detection in a number of samples in which few or no mutations are expected. Radiolabeled probes made from control and test DNAs must be included either together in one experiment, or separately in a series of experiments, to ensure detection of any unreactive T. G mismatches by their complementary A. C mismatch and completely screen a region of DNA for mutations. Unlabeled DNA may also be used, in which case cleavage products can be detected by methods such as silver staining. All DNA is identified in experiments performed in this way, and therefore complete mutation screening is assured. A variety of nucleic acid templates may be used in chemical cleavage. Cloned fragments, polymerase chain reaction (PCR)-amplified fragments, cDNA, and RNA have all been successfully applied to the identification of mutations.
Materials and Reagents Annealing buffer (2 x ): 1.2 M NaCI, 12 mM Tris-HCl (pH 7.5), 14 mM MgCi2; store at room temperature Formamide annealing buffer: 80% formamide, 40 mM piperazine-N, N'bis(2-ethanesulfonic acid) (PIPES) (pH 6.4), 1 mM ethylenediaminetetraacetic acid (EDTA), 0.4 M NaC1; store at - 2 0 ° Hydroxylamine solution: Dissolve 1.39 g hydroxylamine hydrochloride or hydroxylammonium chloride (BDH, Poole, England) in 1.6 ml warm distilled HEO in glass. Add 1.75 ml diethylamine dropwise. Final pH should be 6.0. Store at 4° for up to 2 months Osmium tetroxide buffer (10x): 100 mM Tris-HCl (pH 7.7), I0 mM EDTA, 15% (v/v) pyridine (Sigma, St. Louis, MO); store at -20°C Osmium tetroxide solution: Dissolve contents of a 0.5-g ampoule of osmium tetroxide (Johnson Matthey, Herts, England or Aldrich, Milwaukee, WI) in 12.5 ml distilled H20 in a glass container (not plastic). Stand at room temperature 2-3 days until dissolved. Store well sealed at 4° for up to 3 months. The solution is diluted (1 part in 5 parts)just before use. Should be used in a fume hood
290
MUTAGENESIS AND GENE DISRUPTION
[19]
HOT stop: 0.3 M sodium acetate (pH 5.2), 0.1 mM EDTA, 25/zg/ml tRNA; store at - 2 0 ° Piperidine: Store at room temperature; 10 M stock (Fluka, Ronkonkoma, NY) is diluted (1 part in 10 parts) before use. Should be used in a fume hood Formamide dye: 90% formamide, 20 mM EDTA, 0.3 mg/ml xylene cyanol, 0.3 mg/ml bromphenol blue; store at - 2 0 ° Acryl/bisacrylamide denaturing gels (8%, w/v): 21 g urea, 10 ml 5 × TBE buffer, 40% acrylamide, distilled H20 to 50 ml. Add 250/zl 10% (w/v) ammonium persulfate and 100/zl N,N,N',N'-tetramethylethylenediamine (TEMED) to set the gel TBE buffer (5 × ): 0.45 M Tris, 0.45 M boric acid, 40 mM EDTA; adjust pH to 8.3 with NaOH Acrylamide (40%, w/v): Make 38 g acrylamide and 2 g N,N'-methylenebisacrylamide up to 100 ml in distilled H20. Deionize by mixing with Bio-Rad (Richmond, CA) analytical-grade mixed bed resin AG501X8(D) 20-50 mesh Bind silane solution: Bring 1 liter distilled H20 to pH 3.5 with acetic acid; add 4 ml Bind Silane (LKB, Rockville, MD) and stir for 15 min Silver staining fixing solution: 10% (v/v) ethanol, 5% (v/v) acetic acid; store at room temperature Silver staining solution: 0.011 mol/liter AgNO3 in distilled H20; make fresh Silver staining developer: 0.75 tool/liter NaOH, 0.1 mol/liter formaldehyde solution, 0.0023 mol/liter sodium borohydride; make fresh Silver staining stop solution: 5% (v/v) acetic acid; store at room temperature Methods
General Method Gel Purification for Probe Preparation. Probes may be gel purified from agarose by dialysis 13 or from a 4% (w/v) nondenaturing acrylamide gel, extracting the DNA by soaking the gel slice in 0.6 M sodium acetate overnight. This allows probe fragments to be separated from contaminants, such as PCR artifacts from PCR-amplified probes, other digestion products from probes made by digestion of larger pieces of DNA, and unincorporated nucleotides from labeled probes. 13 j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989.
[19]
CHEMICAL CLEAVAGE OF MISMATCH METHOD
291
DNA-DNA Duplex Formation. Mix test and control DNAs in equal proportions in 1 x annealing buffer to a final volume from 100 to 400/xl for heteroduplex samples. Homoduplex samples are made using one DNA sample only. Cap the tubes with punctured lids. Incubate in a boiling water bath for 5 min, then quench on ice. Briefly centrifuge the tubes to return water drops to the sample volume. Replace the pierced caps with complete caps and incubate at 42 ° for 1 hr. Place the samples on ice. Ethanol precipitate with 0.3 M sodium acetate (pH 5.3) and 70% (v/v) ethanol. Precipitate the DNA at - 20 ° for 30 min, then centrifuge in a microcentrifuge 13,000 rpm at 4 ° for 15-30 min. Wash the samples with ice-cold 70% (v/v) ethanol. Remove all ethanol drops and air dry the pellets for a few minutes. Resuspend the pellets in distilled H=O. Duplex formation is not necessary if the DNA is PCR amplified because heteroduplex molecules will form during the PCR reaction. 14When heteroduplexes are required between a test and a control DNA, the PCR amplification is performed with both DNA samples in one reaction. Hydroxylamine Modification. Mix 6/xl of duplex with 20/zl hydroxylamine solution. It may be necessary to dissolve the crystals in the hydroxylamine solution at 37 ° before adding it to the DNA samples to begin the reaction. Incubate the samples at 37° for up to 3 hr. (Typical time points are 0 min, 30 min, 1 hr, 2 hr, and 3 hr.) The reaction is stopped by the addition of 200/xl HOT stop and 750/zl ethanol. Precipitate the DNA at - 2 0 ° for 30 rain, spin 13,000 rpm at 4° 15-30 min, and wash pellets in icecold 70% (v/v) ethanol. Air dry pellets for a minute. A time course may be performed in pilot work to identify the best reaction conditions. A zero time point should be included by adding HOT stop before adding the hydroxylamine solution. Homozygous control samples should be included in chemical cleavage experiments. Any banding pattern produced can be assigned as background when these control lanes are compared with heteroduplex samples in which mutations are detected. Osmium Tetroxide Modification. Add 2.5 /xl 10 × osmium tetroxide buffer to 6/zl of duplex samples. Add 15 /xl of freshly diluted osmium tetroxide solution to start the reaction. Mix gently with a pipette tip. A yellow precipitate may form. Incubate samples for up to 5 min at 37°. (Typical time points are 0, 1, and 5 min but may be up to 60 rain in some cases.) Stop the reaction with HOT stop and 70% (v/v) ethanol and precipitate the DNA as described for hydroxylamine reactions. Piperidine Cleavage. Resuspend dry pellets in 50/zl of freshly diluted piperidine. Incubate the samples at 90 ° for 30 rain. Transfer the tubes to 14 C. M. N a g a m i n e , K. C h a n , and Y. F. Lau, A m . J. Hum. Genet. 45, 337 (1989).
292
MUTAGENESIS AND GENE DISRUPTION
[19]
ice. Add 50/zl 0.6 M sodium acetate (pH 5.2), 40/zg glycogen, and 300/zl ethanol. Precipitate the DNA as described for hydroxylamine reactions. Resuspend the samples in 5/zl distilled HzO. Add 2/~1 formamide dye. Electrophoresis. Denature the samples at 95 ° for a few minutes or until the sample volume has decreased to 3-4/zl. Quench the tubes on ice. Run the samples alongside size markers on denaturing acrylamide gels [4-8% (w/v) acrylamide] prewarmed to approximately 50 °. After electrophoresis, dry the gels on Whatman (Clifton, N J) 3MM paper. Autoradiography. Dried gels are placed directly against X-ray film. Initial exposures of 10-16 hr are performed, followed by longer exposures as necessary.
Variations of Method
32p Mode. End labeled or internally radiolabeled [32p]dNTP probes may be used in CCM (refs. 3 and 7, respectively); 6000 disintegrations per minute (dpm) of probe is used in each hydroxylamine or osmium tetroxide reaction. Probes up to 2 kb have been used (Fig. 2). 35S Mode. 35S-Labeled dATP probes may also be used in CCM. In this case 50,000 cpm is used per hydroxylamine or osmium tetroxide reaction. 35S-Labeled probes can give better resolution of larger cleavage products. J5 Probes of sufficiently high specific activity can be difficult to obtain with some DNA samples. Probes of up to 2 kb have been used. Unlabeled Mode. Unlabeled DNA may be screened by CCM. Cleavage products are detected by silver staining. This mode of CCM has the advantage of not requiring use of radioisotopes, but is less sensitive than radiolabeled methods in the detection of multiple cleavage products. Cleavage products from 50 to 550 bp can be detected and DNA fragments of up to 600 bp may be screened for mutations. Fifty nanograms of test and 50 ng of control DNA are used to form heteroduplexes and 100 ng of DNA is used to form homoduplexes. 16Reactions are carried out as outlined in the general method. Acrylamide gels are adhered to gel plates to facilitate staining. Before gels are poured the glass plate is soaked in bind silane solution for 1 hr. The plate is rinsed in clean water and air dried. The gels are then poured as usual. After electrophoresis the gel adhered to a glass plate is briefly rinsed in distilled H20, followed by treatment in fixing solution for 30 min. The gel is rinsed in distilled HzO, then soaked for 2 hr in staining solution. Two brief rinses in distilled H20 follow, then 10-20 min in developer. The gel 15 j. A. Saleeba and R. G. H. Cotton, Nucleic Acids Res. 19, 1712 (1991). 16 S. J. R a m o s and R. G. H. Cotton, Hum. Murat. 1, 63 (1992).
[19]
CHEMICAL CLEAVAGE OF MISMATCH METHOD
293
CONTROL MUTANT H O H O 10 60 1 5 10 60 1 5
425 bp 295 bp
t
1
245 bp
I Q
O
-~-- 115 bp
FIG. 2. Autoradiograph showing mutation detection by chemical cleavage using a [32p] dCTP-labeled probe. The control lanes show homoduplex DNA. No cleavage products are seen after reaction with hydroxylamine for 10 or 60 min, or with osmium tetroxide for I or 5 min. Cleavage products are seen with mutant heteroduplex DNA. Hydroxylamine (H) reaction gives products of 295 and 245 bp. Osmium tetroxide (O) reaction gives products of 425, 295, 245, and 115 bp. (Autoradiograph courtesy of S. M. Forrest of the Murdoch Institute.)
is then stored in stop solution. All incubations are performed with gentle rocking. DNA-RNA Duplexes. A useful adaption of CCM allows DNA-RNA duplexes to be screened for mutation sites.17 In this case 5-10/zg of total RNA or 1 /zg poly(A) RNA is annealed with 5 ng probe DNA. Duplexes are formed by mixing DNA and RNA in 1 x formamide annealing buffer, i7 H.-H. M. Dahl, S. R. Lamande, R. G. H. Cotton, and J. F. Bateman, Anal. Biochem. 183, 263 (1989).
294
MUTAGENESIS AND GENE DISRUPTION
[19]
incubating at 80 ° for 5 min, then at 48 ° for 1 hr. Samples are then precipitated and reactions are performed as described in the general method. Hybond Method. Gogos et al. 18have published an adaptation of CCM whereby probe DNA is hybridized to DNA applied to a membrane. This allows time-consuming ethanol precipitation steps to be simplified. DNA is transferred from one reaction to the next by lifting the membrane between tubes.
Concluding Remarks and Discussion of Problems The CCM method is based on chemical and not enzymatic reactions, and therefore the problems often encountered in molecular biology, for example enzyme reactions that are sensitive to contaminants such as those present in some DNA preparations, do not occur here. However, care must be taken with CCM in correct solution preparation. Overreaction by piperidine is sometimes seen in chemical cleavage. This is indicated by the presence of a ladder of cleavage products at each base position. Repeat the reaction with reduced piperidine reaction time, and check that the piperidine stock is diluted 1 in 10 before use. Lack of mutation detection may be due to a failure to form heteroduplex molecules. Heteroduplexes may be destroyed after formation by heating samples to dissolve DNA pellets. Alternatively, a lack of reaction may be due to aging chemical solutions. These should be remade. Lack of hydroxylamine and osmium tetroxide modification of mismatches is identified by failure of the probe to decrease in intensity with increased incubation time in the presence of these chemicals. In the case of diploid organisms, heteroduplex molecules will be formed if selfoDNA is heated and annealed, when a heterozygous mutation is present. The mutation can then be identified by CCM. If a homozygous mutation is present, CCM with self-DNA will not identify the mutation. In this way heterozygous and homozygous mutations can be distinguished. ~9 Screening more than 2-kb stretches of DNA for mutations may require that a series of probes be designed. These should be arranged so that each probe will overlap neighboring probes by 20-30 bp. Mutations at the ends of probes may be missed by CCM because breathing of the duplex can occur at fragment ends. H Probes obtained by PCR amplification will also Is j. A. Gogos, M. Karayiorgou, H. Aburatani, and F. C. Kafatos, Nucleic Acids Res. 18, 6807 (1990). ~9 I. Dianzani, S. M. Forrest, C. Camaschella, E. Gottardi, and R. G. H. Cotton, Am. J. Hum. Genet. 48, 423 (1991).
[20]
RAPID INTRODUCTION OF RESTRICTION SITES
295
lead to mutations being missed at the ends of probe fragments because regions covered by primers will have primer sequence and not the sequence of mutations, if they were originally present at those sites before amplification. In conclusion, chemical cleavage offers a reliable screening method for mutation detection.
[20] I n t r o d u c i n g R e s t r i c t i o n Sites i n t o D o u b l e - S t r a n d e d Plasmid DNA
By
D A V I D C . K A S L O W a n d DAVID J. RAWLINGS
General Introduction Double-stranded DNA plasmids are the workhorses of molecular biology, especially for cloning foreign genes and expressing recombinant protein. Although a number of pBR322/pUC-derived plasmids now contain a large choice of restriction enzyme sites, many of the specialized vectors for prokaryotic or eukaryotic expression have only a few restriction sites in the cloning area. Oftentimes these sites are incompatible with the fragment to be cloned, or in a case such as ours, a blunt end site and a site for an enzyme that cuts infrequently in the target DNA are not present but are required for high-efficiency blunt end ligation.t The classical approach to introducing new restriction sites is to construct a pair of synthetic oligonucleotides that anneal to one another to yield a double-stranded product that contains single-stranded overhangs at each end complementary to the restriction sites in the vector. This approach is necessary if the restriction site(s) in the vector produce two 5' overhangs, two 3' overhangs, or if one of the sites is blunt ended, if, on the other hand, one of the sites results in a 5' overhang, and the other site has a 3' overhang, then conceivably a single-stranded oligonucleotide could anneal to each end (Fig. 1). This would leave a single-stranded gap in the plasmid that on transformation is filled in by host (bacterial) DNA repair mechanisms. Principle of Method The following is a rapid method for introducing new restriction enzyme sites or any short DNA sequence of choice into virtually any doublei p. Upcroft and A. Healey, Gene 51, 69 (1987).
METHODS IN ENZYMOLOGY,VOL. 217
296
MUTAGENESIS AND GENE DISRUPTION
[20]
A 5 'P- IGATCCCGCGGAGATCTATCGATCTGCA 13 '
3 ' IAGCT
B 123456M
m
I
il
II
FIG. 1. (A) Schematic depiction of the introduction of a single-stranded oligonucleotide into a double-stranded plasmid. Shown are the complete nucleotide sequence of kinased 27mer (box), protruding single-stranded ends of BamHI/PstI-digested pUC19 (open ended boxes), new Bglll site in 27-mer (bold), and 5' phosphate removed from protruding 5' end of plasmid by bacterial alkaline phosphatase (italics). (B) Ethidium bromide-stained 1% (w/ v) agarose-TAE gel. BgllI digest of plasmid DNA from nonhybridizing colony (lane 1) and hybridizing colonies (lanes 2-6) of colony lift probed with 27-mer (Table I). M, Molecular weight markers; CC, closed circular (uncut); L, linear (single cut).
stranded plasmid containing at least one site that produces a 5' overhang and one that produces a 3' overhang. A single-stranded synthetic oligonucleotide containing the new restriction sites is constructed such that one end contains the complementary sequence to the 5' overhang and the other end contains the complementary sequence to the 3' overhang (Fig. 1). As each end of the single-stranded oligonucleotide can anneal to a complementary sequence in the plasmid, the cost and time of producing a second oligonucleotide are avoided. Since first describing this method, 2 we have found that dephosphorylating the 5' overhang in the vector increases the relative efficiency of 2 D. C. Kaslow and D. J. Rawlings, Nucleic Acids Res. 17, 10135 (1989).
[20l
RAPID INTRODUCTION OF RESTRICTION SITES
297
the technique by reducing the background, that is, absolute number of nonrecombinant clones. This background is due to the presence of (1) uncut vector, (2) religation of the vector digested with only the 3' overhang restriction enzyme or with only the 5' overhang restriction enzyme (incomplete digestion by the other restriction enzyme), or (3) intramolecular ligation of the vector by the 5' overhang ligating with the 3' overhang to yield a single-stranded blunt end ligation product. Based on the requirement for a 5'-phosphate for efficient ligation, we reasoned that the latter two undesired products could be significantly decreased by removing the 5'-phosphate present on the 5' overhang (Fig. 1A, italics) of the vector. We further reasoned that because Escherichia coli alkaline phosphatase preferentially hydrolyzes the 5'phosphate from protruding 5' ends of double-stranded DNA at 37 ° but inefficiently removes the 5'-phosphate from recessed 5' ends at that temperature (temperature optimum for blunt or recessed 5' ends is 60°), 3 bacterial alkaline phosphatase (BAP) treatment of the doubly digested vector would mainly result in dephosphorylation of the unused 5' overhang (Fig. 1A, italics). Experimentally, we find that BAP treatment of the doubly digested vector reduces the number of nonrecombinant plasmids in the ligation. Materials, Reagents, and Their Use Oligonucleotides: 5'-GATCCTGCA-3' (9-mer), 5'-CGATGTCGACC GC-Y (13-mer), 5'-CGATGTCGACTGCA-3' (14-mer), 5'-GATC CATCGATCTGCA-Y (16-mer), 5'-GATCCCGCGGAGATCTATC GATCTGCA-3' (27-mer), 5'-GATCCGCGCGCGGCCGCGATAT CGCAGCTGAAGCTTGCA-3' (39-mer, nucleotides annealing to vector in bold). Following synthesis on an Applied Biosystems (Foster City, CA) DNA synthesizer, the oligonucleotides were cleaved from the resin in 1 ml of concentrated ammonium hydroxide by incubation at room temperature for I hr, deprotected by incubation for 4 to 24 hr at 55 °, and lyophilized in a speed vacuum centrifuge. The resulting crude oligonucleotides were resuspended in 1 x TE [10 mM Tris-HCl (pH 8.0), 1 mM ethylenediaminetetraacetic acid (EDTA)] to a final concentration of 1.5/~g/ml, as determined by absorbance at 260 nm, and used without any further purification Vector DNA: pUC19 supercoiled DNA was prepared from saturated cultures of transformed DH10B E. coli by the Qiagen (Qiagen, Inc., Chatsworth, CA) method, and resuspended in 1 x TE at a final concentration of 1 /xg/~l 3 F. Cobianchi and S. H. Wilson, this series, Vol. 152, p, 99.
298
MUTAGENESIS AND GENE DISRUPTION
[20]
T4 Polynucleotide kinase, 5 x forward reaction buffer: 300 mM TrisHC1 (pH 7.5), 75 mM 2-mercaptoethanol, 50 mM MgCI2, 1.65/zM ATP T4 DNA ligase, 5 x ligase buffer: 0.25 M Tris-HCl (pH 7.6), 50 mM MgCI 2 , 5 mM ATP, 5 mM dithiothreitol (DTT), 25% (w/v) polyethylene glycol (PEG) 8000 Restriction enzymes and buffers: Purchased from Bethesda Research Laboratories (Gaithersburg, MD) Electroporation competent cells: Electrocompetent DH10B E. coli cells can be purchased (Bethesda Research Laboratories) or prepared as follows (Dr. J. Jessee, personal communication, 1992). Host bacterial cells, such as DH10B, are grown to saturation overnight in 5 ml of LB broth at 37°. One milliliter of the overnight growth is diluted with 100 ml of SOB [2% (w/v) tryptone (Difco, Detroit, MI), 0.5% (w/v) yeast extract (Difco), 10 mM NaCI] and allowed to grow at 37° to midto late log phase (OD600 of 0.5 to 0.8). The bacterial pellet collected by centrifugation at 3000 g for 15 min is washed twice in 100 ml of 10% (v/v) glycerol in high-performance liquid chromatography (HPLC)-grade water (the water used in this wash must be of the highest puritymwe use Baxter HPLC water, Baxter Healthcare Corp., Muskegon, MI). The final pellet is resuspended in 10% (v/v) glycerol in HPLC-grade water to an optical density at 600 nm of 250, and stored at - 7 0 ° in 25-t~1 aliquots. The diluted ligation mix (see below) is added to 20 t~l of electrocompetent cells that have been thawed on ice. The transformation mix is transferred to a 0.1-mm cuvette (Bio-Rad, Richmond, CA) and electroporated at 1.8 mV, 25 /zF, 200 1~ (Bio-Rad). The sample is immediately transferred to 1 ml of SOC [2% (w/v) tryptone (Difco), 0.5% (w/v) yeast extract (Difco), 10 mM NaCI, 2.5 mM KC1, 10 mM MgC12, 10 mM MgSO 4, 20 mM glucose], and then incubated for 1 hr at 37°. One, 10, 100, and 400 tzl of the medium are plated onto LB agar-ampicillin (50/zg/ml) plates, and incubated overnight at 37° Filters for hybridization: Bacterial colonies on LB agar-ampicillin plates are transferred to nylon membranes by placing the dry, sterile filters directly on the plates. The filters are punctured asymmetrically with a 23-gauge needle, then transferred from the plates to 1-ml pools of denaturing buffer (0.5 N NaOH, 1.5 M NaC1) on plastic wrap. After denaturation for several minutes, the filters are briefly air dried on Whatman (Clifton, N J) paper. The denaturation process is repeated twice again. The filters are placed in 1-ml pools of neutralizing buffer (1 M ammonium acetate, 20 mN NaOH) for several minutes, baked for 30-60 min at 80 °, washed [6 × SSC (1 × SSC is 0.15 M NaC1, 0.015 M sodium citrate),
