Gene, 109 (1991) 115-119 © 1991 Elsevier Science Publishers B.V. All rights reserved. 0378-1119/91/$03.50
115
GENE 06194
Short Communications Optimised cDNA size selection and cloning procedure for the construction of representative plasmid cDNA libraries (Sucrose gradient; drop dialysis; E. coli; electroporation; ligase; pCDM8; primary transformants; insert size; expression library)
B.L. Kieffer Ecole Sup$rieure de Biotechnologie, 67085 Strasbourg (France)
Received by J.-P. Lecocq: 19 April 1991 Revised/Accepted: 29 July/19 September 1991 Received at publishers: 23 September 1991
SUMMARY Plasmid libraries are more versatile than phage libraries since they allow expression cloning in eukaryotic systems. However, high numbers of primary clones are sometimes difficult to obtain and more efficient transformation procedures are often required. In this paper, a detailed protocol is presented, for the construction of large plasmid libraries from ng quantities of eDNA, based on a highly efficient transformation step. Drop dialysis and electroporation are optimised: complete ligase removal and yeast tRNA addition before dialysis appear critical, while exclusive use of double-distilled water as well as rapid preparation of fresh cells provide excellent electroporation yields. A novel, simple and flexible eDNA size selection procedure is also presented, based on sucrose density gradient. Two libraries were constructed using an expression vector for mammalian cells (pCDM8) and its Escherichia coli host strain (MC1061[p3]). Numbers of 2 to 10 × 106 primary transformants were obtained from 1 #g of poly(A) + RNA. Up to 85 % of the clones had inserts and half of the inserts were larger than 1.5 kb.
INTRODUCTION
Cloning a rare protein requires the screening of large libraries. Phage Ais the vector of choice for the construction of such libraries since concatemer ligafion and in vitro Correspondenceto: Dr. B.L. Kieffer, Ecole Sup6rieurede Biotechnologie, 11 rue Humann, 67085Strasbourg Cedex (France) Tel. 88358727; Fax 88363828.
Abbreviations: Ap, ampicillin;bp, base pair(s); BSA,bovineserum albumiz~;cDNA, DNA complementaryto RNA; cfu, colony-formingunit(s); DTT, dithiothreitol; EtdBr, ethidium bromide; HS, heat shock; kb, kilobase(s) or 1000bp; oligo, oligodeoxyribonucleotide; p, plasmid; Pol I, polymerase I; poly(A)+RNA, polyadenylated RNA; R resistance; rt, room temperature; SOC, Bacto tryptone 2%/yeast extract 0.5%/NaCI 10 mM/KCI2.5 mM/MgCl210 mM/MgSO4 10 mM/glucose20 raM; Tc, tetracycline; TE, 10 mM Tris. HCi pH 8/1 mM EDTA; u, unit(s); [ ], denotes plasmid-carrier state.
packaging are highly efficient. A million recombinants are therefore easily obtained from 1 #g of mRNA (Sambrook et al., 1989). However, more versatile cloning strategies are available when using plasmids as vectors (Kimmel and Berger, 1987), and plasmid libraries offer additional possibilities, such as scree,ing for expression in eukaryotic cells. Although several strategies have been developed to improve eDNA ligation into plasmid vectors, a strong limitation lies in poor yields when transferring this DNA to E. coil cells. H S transformation of competent bacteria (Hanahan, 1986) rarely exceeds 10s cfu/#g vector efficiencies and is strain-dependent, thus the generation of more than 100000 clones, using this procedure, remains a challenge (Helfman et al., 1987). Electrotransformation is more efficient for many E. coil strains (109 to 10~° cfu/#g of supercoiled DNA; see Dower et al.. 1988) and is comparable to phage Ain vitro packaging
116 yields. Unfortunately, the electroporation efficiency drops about 1000-fold when salts are present in the sampl~ (Jacobs et al., 1990) and a dialysis step has to be performed when a ligation mixture is used. Electroporation efficiency is then hampered by poor dialysis recovery if small amounts of D N A are available and the overall process is then as inefficient as HS transformation. For the construction of a c D N A library in p C D M 8 , an expression vector for COS cells (Aruffo and Seed, 1987), I optimised the electrotransformation o f a ligation reaction mixture into the E. coli host strain MC1061[p3] and succeeded in getting final overall yields exceeding phage A in vitro packaging efficiencies. The high quality of a library depends not only on the number of clones but also on the size of the inserts. A stringent size selection ofthe e D N A before cloning is necessary for a high ratio of full-length clones. The gel-filtration method described in Sambrook et al. (1989) eliminates the excess of linkers and very small e D N A molecules ( < 500 bp) but does not allow real size selection among large fragments. Salt gradients are difficult to perform and agarose gels are better avoided because of ligase-inhibiting contaminants, although the purity of commercially available agarose has improved. Sucrose gradients are known to generate highly pure D N A but to our knowledge, are only used for preparation of single D N A fragments or for genomic DNA fractionation (Ausubel et ai., 1987). I adapted and simplified the sucrose gradient method for e D N A fraetionation. The procedure is easy to perform (no gradient is poured), flexible (size selection range 0.2-10 kb) and provides high quality D N A for ligation. In this paper, critical parameters are described for maximum electroporation efficiency. Then, a novel, convenient and highly efficient protocol is presented and discussed for e D N A size fractionation/ligation into vector/transformation into E. coll. Using this procedure, I routinely obtain more than 2 × 106 independent clones (insert size > 1.5 kb for half of them) out of 1 #g of poly(A) ÷ RNA.
EXPERIMENTAL AND DISCUSSION (a) Eleetroporation of a ligation mixture: optimisation
H) Cell preparation Water quality and rapid handling are important. The MCI061[p3] cells are MCI061 cells (Hanahan, 1986) which contained a stably maintained single-copy 57-kb plasmid (p3). This plasmid carries amber-mutated Ap R and T c R elements and allows the selection of SupF plasmids (CDM8 in this case). MC1061 cells are poorly transformed by the HS procedure (5 × 10v-108//~g DNA). Electroporation is more efficient since a 109 cfu//~g D N A transfor-
mation yield was obtained using the protocol described by Dower et al. (1988). I found that this yield could be improved to 2 - 6 x 10 ~° cfu/#g D N A by considering the following. First, for best and reproducible efficiency, a fresh preparation of cells should be used rather than stored cells. Also, electroporation should be performed as soon as the cells are ready since a two- and threefold drop in efficiency was observed when cells were left on ice for 1 and 2 h, respectively, before use (data not shown). Second, the purity of the water used for cell prepaxation and sample dialysis is critical and should be tested. In the present work, five times more transformants were obtained when doubledistilled water was used instead of deionised water (e.g., MilliQ system, Millipore) (data not shown).
(2) Sample preparation The ligation mixture has to be treated before dialysis (Table I). CsCl-purified piasmid (pCDM8), added to a standard ligation mixture, was tested for electroporation efficiency. Efficiency of 100% was defined as the number of transformants obtained by direct electroporation of the plasmid added to water. Table I shows that dialysis of the ligation mixture was reasonably efficient (25 %) when small TABLE I Overall transformation yield (~) depending on the ligation mixture" DNA amounff Ligaseh
Ligationmixture before dialysis
! ng
10 ng
100 ng
I: 6 diluted
No treatmentd + Yeast tRNA~ Phenol extr. + tRNA f
0.2 6.5 34.4
9.9 18.4 36.4
25.0 34.8 27.3
Undiluted
No treatment° + Yeast tRNA~ Phenol extr. + tRNA r
...... 1.8
115
GENE 06194
Short Communications Optimised cDNA size selection and cloning procedure for the construction of representative plasmid cDNA libraries (Sucrose gradient; drop dialysis; E. coli; electroporation; ligase; pCDM8; primary transformants; insert size; expression library)
B.L. Kieffer Ecole Sup$rieure de Biotechnologie, 67085 Strasbourg (France)
Received by J.-P. Lecocq: 19 April 1991 Revised/Accepted: 29 July/19 September 1991 Received at publishers: 23 September 1991
SUMMARY Plasmid libraries are more versatile than phage libraries since they allow expression cloning in eukaryotic systems. However, high numbers of primary clones are sometimes difficult to obtain and more efficient transformation procedures are often required. In this paper, a detailed protocol is presented, for the construction of large plasmid libraries from ng quantities of eDNA, based on a highly efficient transformation step. Drop dialysis and electroporation are optimised: complete ligase removal and yeast tRNA addition before dialysis appear critical, while exclusive use of double-distilled water as well as rapid preparation of fresh cells provide excellent electroporation yields. A novel, simple and flexible eDNA size selection procedure is also presented, based on sucrose density gradient. Two libraries were constructed using an expression vector for mammalian cells (pCDM8) and its Escherichia coli host strain (MC1061[p3]). Numbers of 2 to 10 × 106 primary transformants were obtained from 1 #g of poly(A) + RNA. Up to 85 % of the clones had inserts and half of the inserts were larger than 1.5 kb.
INTRODUCTION
Cloning a rare protein requires the screening of large libraries. Phage Ais the vector of choice for the construction of such libraries since concatemer ligafion and in vitro Correspondenceto: Dr. B.L. Kieffer, Ecole Sup6rieurede Biotechnologie, 11 rue Humann, 67085Strasbourg Cedex (France) Tel. 88358727; Fax 88363828.
Abbreviations: Ap, ampicillin;bp, base pair(s); BSA,bovineserum albumiz~;cDNA, DNA complementaryto RNA; cfu, colony-formingunit(s); DTT, dithiothreitol; EtdBr, ethidium bromide; HS, heat shock; kb, kilobase(s) or 1000bp; oligo, oligodeoxyribonucleotide; p, plasmid; Pol I, polymerase I; poly(A)+RNA, polyadenylated RNA; R resistance; rt, room temperature; SOC, Bacto tryptone 2%/yeast extract 0.5%/NaCI 10 mM/KCI2.5 mM/MgCl210 mM/MgSO4 10 mM/glucose20 raM; Tc, tetracycline; TE, 10 mM Tris. HCi pH 8/1 mM EDTA; u, unit(s); [ ], denotes plasmid-carrier state.
packaging are highly efficient. A million recombinants are therefore easily obtained from 1 #g of mRNA (Sambrook et al., 1989). However, more versatile cloning strategies are available when using plasmids as vectors (Kimmel and Berger, 1987), and plasmid libraries offer additional possibilities, such as scree,ing for expression in eukaryotic cells. Although several strategies have been developed to improve eDNA ligation into plasmid vectors, a strong limitation lies in poor yields when transferring this DNA to E. coil cells. H S transformation of competent bacteria (Hanahan, 1986) rarely exceeds 10s cfu/#g vector efficiencies and is strain-dependent, thus the generation of more than 100000 clones, using this procedure, remains a challenge (Helfman et al., 1987). Electrotransformation is more efficient for many E. coil strains (109 to 10~° cfu/#g of supercoiled DNA; see Dower et al.. 1988) and is comparable to phage Ain vitro packaging
116 yields. Unfortunately, the electroporation efficiency drops about 1000-fold when salts are present in the sampl~ (Jacobs et al., 1990) and a dialysis step has to be performed when a ligation mixture is used. Electroporation efficiency is then hampered by poor dialysis recovery if small amounts of D N A are available and the overall process is then as inefficient as HS transformation. For the construction of a c D N A library in p C D M 8 , an expression vector for COS cells (Aruffo and Seed, 1987), I optimised the electrotransformation o f a ligation reaction mixture into the E. coli host strain MC1061[p3] and succeeded in getting final overall yields exceeding phage A in vitro packaging efficiencies. The high quality of a library depends not only on the number of clones but also on the size of the inserts. A stringent size selection ofthe e D N A before cloning is necessary for a high ratio of full-length clones. The gel-filtration method described in Sambrook et al. (1989) eliminates the excess of linkers and very small e D N A molecules ( < 500 bp) but does not allow real size selection among large fragments. Salt gradients are difficult to perform and agarose gels are better avoided because of ligase-inhibiting contaminants, although the purity of commercially available agarose has improved. Sucrose gradients are known to generate highly pure D N A but to our knowledge, are only used for preparation of single D N A fragments or for genomic DNA fractionation (Ausubel et ai., 1987). I adapted and simplified the sucrose gradient method for e D N A fraetionation. The procedure is easy to perform (no gradient is poured), flexible (size selection range 0.2-10 kb) and provides high quality D N A for ligation. In this paper, critical parameters are described for maximum electroporation efficiency. Then, a novel, convenient and highly efficient protocol is presented and discussed for e D N A size fractionation/ligation into vector/transformation into E. coll. Using this procedure, I routinely obtain more than 2 × 106 independent clones (insert size > 1.5 kb for half of them) out of 1 #g of poly(A) ÷ RNA.
EXPERIMENTAL AND DISCUSSION (a) Eleetroporation of a ligation mixture: optimisation
H) Cell preparation Water quality and rapid handling are important. The MCI061[p3] cells are MCI061 cells (Hanahan, 1986) which contained a stably maintained single-copy 57-kb plasmid (p3). This plasmid carries amber-mutated Ap R and T c R elements and allows the selection of SupF plasmids (CDM8 in this case). MC1061 cells are poorly transformed by the HS procedure (5 × 10v-108//~g DNA). Electroporation is more efficient since a 109 cfu//~g D N A transfor-
mation yield was obtained using the protocol described by Dower et al. (1988). I found that this yield could be improved to 2 - 6 x 10 ~° cfu/#g D N A by considering the following. First, for best and reproducible efficiency, a fresh preparation of cells should be used rather than stored cells. Also, electroporation should be performed as soon as the cells are ready since a two- and threefold drop in efficiency was observed when cells were left on ice for 1 and 2 h, respectively, before use (data not shown). Second, the purity of the water used for cell prepaxation and sample dialysis is critical and should be tested. In the present work, five times more transformants were obtained when doubledistilled water was used instead of deionised water (e.g., MilliQ system, Millipore) (data not shown).
(2) Sample preparation The ligation mixture has to be treated before dialysis (Table I). CsCl-purified piasmid (pCDM8), added to a standard ligation mixture, was tested for electroporation efficiency. Efficiency of 100% was defined as the number of transformants obtained by direct electroporation of the plasmid added to water. Table I shows that dialysis of the ligation mixture was reasonably efficient (25 %) when small TABLE I Overall transformation yield (~) depending on the ligation mixture" DNA amounff Ligaseh
Ligationmixture before dialysis
! ng
10 ng
100 ng
I: 6 diluted
No treatmentd + Yeast tRNA~ Phenol extr. + tRNA f
0.2 6.5 34.4
9.9 18.4 36.4
25.0 34.8 27.3
Undiluted
No treatment° + Yeast tRNA~ Phenol extr. + tRNA r
...... 1.8
