Combinatorial libraries by cassette mutagenesis

1834-1835

Nucleic Acids Research, 1995, Vol. 23, No. 10

© 1995 Oxford University Press

Combinatorial libraries by cassette mutagenesis Belen Borrego1, Anja Wienecke1 and Andreas Schwienhorst* Department of Molecular Evolution Biology, Institute for Molecular Biotechnology, BeutenbergstraBe 11, PO Box 100813, D-07745 Jena, Germany and department of Biochemical Kinetics, Max-Planck-Institute for Biophysical Chemistry, Gottingen, Germany Received February 13, 1995; Revised and Accepted March 31, 1995

In recent years random mutagenesis in line with a strong means of selection has proved to be a valuable tool for identifying structural determinants of functional molecules such as important nucleotides in bacterial promoters (1,2) or essential amino acid residues located in active sites of various enzymes (3-9). Random sequences are usually introduced on the level of double-stranded DNA by ligating chemically synthesized, partially randomized oligo-DNAs into a vector, previously digested with appropriate restriction enzymes, thus replacing a part of the original sequence between the two restriction sites by partially randomized sequences. Therefore, the application of random cassette mutagenesis is limitted to DNA sequences that already contain appropriate cleavage sites for (rare) restriction enzymes. Certainly, new and useful cleavage sites can also be engineered into the target sequence as far as the sequences introduced do not interfere with the activity of interest Alternatively, there are examples where partially randomized sequences may also be introduced by a two-step site-directed mutagenesis procedure (5). Here we propose a novel method of random cassette mutagenesis, which is capable of introducing random sequences exclusively without altering adjacent sequences, for example by engineering cleavage sites for restriction sites that remain in the final product of cloning. In a first step therecognitionsite for Bcg\ (wirjN-CGA-N6-TGC-N|2u) is introduced into the target sequence by standard site directed mutagenesis (10-13). At present, Bcgl is the only restriction enzyme available that cleaves DNA substrates twice, so as to excise its recognition site such that cleavage would yield two different cohesive two-base-overhangs. Hence cleavage sites are arbitrary in sequence, they can be generated in any sequence of choice, given that the recognition site is engineered in the appropriate distance to the cleavage sites. Cleavage then will reveal a digested target sequence with a piece of typically 34 bp missing, that could be replaced by partially randomized sequences. The combinatorial library prepared in this way can be subjected to any kind of selection strategy. In a first selection experiment the sequence of the Tetrahymena intron obtained by PCR on the plasmid pT7TTlA3 (14) using primer Intro 1 (GCT GGG TAC CAC TCT CTA A AT AGC) and Intro2 (CTT ACG AGT ACT CCA AAA CTA ATC AAT ATA C) was digested with Kpnl and Seal and ligated into the KpnVScal double digested vector 'pAS-R' obtained by ligating the complementary oligo-DNAs Adaptor 1 (CTA GAA GTA CTC GAG

* To whom correspondence should be addressed

CAG CT) and Adaptor 2 (GCT CGA GTA CTT) into pBluescript KS+ (Stratagene, La Jolla, CA) double digested with Kpnl and Sacl, yielding pAS-Intron which had the lacZgene interruptedly the Tetrahymena intron. Escherichia coli strain XL 1-Blue carrying pAS-Intron proved to be fully active in exhibiting galactosidase activity in vivo indicating that the intron is successfully spliced in vivo. Using a variant of the site-directed mutagenesis protocol of Hall et al. (13) a Bcgl site was introduced at a defined position of the intron yielding vector pAS-Bcgl which proved to be inactive in galactosidase activity in vivo (Fig. 1). Finally, the DNA oligomers Libl (GAT ATG GAN NNN NNT CAC A) and Lib2 (TGA NNN NNN TCC ATA TCA A), containing six randomized positions each (N meaning 25% each of the four nucleotides G, A, C and T), were ligated into vector pAS-Bcgl previously linearized with Bcgl, using 1/5 j-value concentration (15) of vector and 11-fold excess of insert Transformants were plated on agar plates containing 20 |Xg/ml ampicillin, 10 (ig/ml tetracyclin and 40 (0.1 X-gal (20 mg/ml in dimethylformamide) and 4 u:l IPTG (200 mg/ml) previously spread on the surface of the agar plates. Altogether different experiments revealed an average transformation efficiency of typically 1-2.2 x 103 transformants/(ig of linearized vector using the protocol of Crouse et al. (16). Phagemid DNA from bright blue colonies as well as from clear white colonies was prepared and subjected to automated sequencing using an ABI 373A sequencer. Sequencing revealed that the cassette mutagenesis technique presented here is capable of introducing stretches of random sequence into a given target sequence without adding other non-native sequences. No religated vector could be traced among 43 species sequenced so far. Results indicate that the method presented here clearly has the potential to easily generate combinatorial libraries that can be subjected to selection procedures, for example, to determine structural determinants of functional molecules such as ribozymes.

ACKNOWLEDGEMENTS This work was supported by grant Schw578/1-1 from the Deutsche Forschungsgemeinschaft (DFG). Vector pT7TTlA3 was a kind gift from Tan Inoue at the Salk Insitute (La Jolla).

Nucleic Acids Research, 1995, Vol. 23, No. 10 1835 »•

391

8C*I

(S*aX)

GGTACCACTCTCTAAATA

CTGTTGATATGGATGCAGTTCACAGACTA

AGTACTCGAGCAGCTC

CCATGGTGAGAGATTTAT

, GACAACTATACCTACGTCAAGTGTCTGAT .

TCATGAGCTCGTCGAG

pAS-Intron

Site-directed

mutagenesis \

BagX

BagZ

recognition mlf

CTGTipATATGQATC 3CAACT, GACKACTATACCTAG MTTGA 1

ATCCATTTCJ^ZAGACTA TAGGTAAAGAGTCTGAT

J

pAS-Bcgl

Library cloning

CTGTT GATATGOAMMOOOTTCACA GACTA GAC AACTATACl IHIMMIWAGT GTCTGAT

pJkS-ULbl

Figure 1. Steps of the novel cassette mutagenesis technique used for library design. The Tetrakymena group I intron sequence (position 1-414) was cloned into the multiple cloning site of the LacZ gene in pBluescript II KS+ such that splicing of the intron would reveal a fully active lacZ gene product, thus providing a-complementation for blue/white color selection of recombinant phagemids. Using standard site directed mutagenesis methods a Bcgl site was first introduced into the intron sequence of pAS-intron. The resulting vector pAS-Bcgl was digested with Bcgl and the oligonucleondes Libl and Lib2 representing a library with six randomized positions (positions 253-258) were ligated into the linearized vector replacing the missing intron sequences.

REFERENCES 1 Horwitz, M.S.Z., Loeb, LA. (1986) Proc. NatL Acad. ScL USA 83, 7405-7409. 2 Horwitz, M.S.Z., Loeb, L.A. (1988) J. BioL Chem. 263, 14724-14731. 3 Reidhaar-Olson, J.F., Sauer, R.T. (1988) Science 241, 53-57. 4 Kaiser, C.A., Preuss, D., Grisafi, P., Botstein, D. (1987) Science 235, 312-317. ' 5 Dunn, I.S., Cowan, R., Jennings, PA. (1988) Prot. Engng. 2, 283-291. 6 Dube, D.K., Loeb, LA. (1989) Biochemistry 28, 5703-5707. 7 Oliphant, A.R., Struhl, K. (1989) Proc. NatL Acad Sci. USA 86, 9094-9098. 8 Hermes, J.D., Blacklow, S.C., Knowles, J.R. (1990) Proc. NatL Acad. ScL USA 87, 696-700.

Munir, ICM., French, D.C., Dube, D.K., Loeb, L.A. (1992) J. BioL Chem. 267, 6584-6589. 10 Zoller, MJ., Smith, M. (1982) Nucleic Acids Res. 10, 6487-6500. 11 Higuchi, R., Krummel, B., Saiki, R.K. (1988) Nucleic Acids Res. 16, 7351-7367. 12 Ho, S.N., Hunt, H£>., Horton, R^l., PuUen, J.K., Pease, L.R. (1989) Gene 77,51-59. 13 Hall, L., Emery, D.C. (1991) Prot. Engng. 4, 601. 14 Zaug, AJ., Been, M.D. and Cech, T. (1986) Nature 324, 429-433. 15 Dugaiczyk, A., Boyer, H.W., Goodman, H.M. (1975) J:MoL BioL 96, 171-184. 16 Crouse, G.F., Frischauf, A., Lehrach, H. (1983) Methods EnzymoL-101, 78-90.

9