PEDS Advance Access originally published online on January 13, 2006
Protein Engineering Design and Selection 2006 19(3):135-140; doi:10.1093/protein/gzj008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
SHORT COMMUNICATION |
Frame shuffling: a novel method for in vitro protein evolution
1Department of Protein Engineering, Cancer Institute, Japanese Foundation for Cancer Research, 3-10-6, Ariake, Koto-ku, Tokyo 135-8550, 2CREST, Japanese Science and Technology Corporation Agency and 3Biophysical Chemistry Laboratory, Institute of Physical and Chemical Research (RIKEN), RIKEN Harima Institute, Mikazuki-cho, Sayo, Hyogo 679-5143, Japan
4 To whom correspondence should be addressed.E-mail: kshiba{at}jfcr.or.jp
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionAcknowledgementsReferences |
|---|
We describe ‘frame shuffling’, a novel method for preparing artificial protein libraries. With this method, a Y-family DNA polymerase known to introduce frame shift mutations at high rates is utilized to scramble the reading frames of a parental gene. The resultant progeny produce mutant proteins having segmental sequence changes. Such frame-shuffled mutant proteins exhibit physicochemical properties that differ from those of proteins obtained using conventional mutagenesis.
Keywords: artificial protein/frame shift mutation/in vitro evolution/microgene/Y-family DNA polymerase
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionAcknowledgementsReferences |
|---|
Efficient mutagenesis is critical for the construction of a diversity library and is therefore the key to successful evolution of proteins in vitro (Neylon, 2004
). For that reason, considerable effort has gone into characterizing mutant DNA polymerases and seeking replication conditions that efficiently introduce DNA base substitutions (Cirino et al., 2003). By contrast, mutants or conditions that provoke frame shift mutations are highly undesirable because they usually allow the appearance of a termination codon, which leads to the production of a defective library with immature open reading frames. We have been working on a ‘microgene’-based protein creation system in which blocks of microgenes are polymerized to make artificial protein libraries having long reading frames (Shiba et al., 1997, 2002a,b, Kashiwagi and Shiba, 2004; Saito et al., 2004; Shiba, 2004). With this microgene strategy, we utilize all three (or six) (Kashiwagi and Shiba, 2004) reading frames inherent within a single microgene, as it is the combinatorial linkage of these frames that lends diversity to the polymers. We designed the microgenes so that they do not contain termination codons in any of their reading frames; furthermore, by taking advantage of the degeneracy of the genetic code, certain functions or structures can be rationally encrypted into two or three reading frames at the same time (Saito et al., 2004; Shiba, 2004). Thus, given this approach to achieving diversity, frame shift mutations are preferable to sparse sampling of protein sequence space (Figure 1A).
|
Y-family DNA polymerase is an evolutionally conserved and specialized DNA polymerase (Ohmori et al., 2001
) that substitutes for high-fidelity polymerase at times of replication crisis (Goodman and Tippin, 2000; Friedberg et al., 2002). In vivo, this enzyme has the ability to synthesize DNA past a variety of DNA lesions (e.g. abasic defects) (Boudsocq et al., 2001); in vitro, it catalyzes high-error-rate, low-fidelity DNA synthesis with undamaged templates (Wagner et al., 1999; Ohashi et al., 2000; Strauss et al., 2000; Zhang et al., 2000; Kokoska et al., 2002). Interestingly, because of its weak association with the DNA substrate (Boudsocq et al., 2004), Y-family DNA polymerase frequently generates frame shift mutations both in vitro (frame shift rates are reported to be 8 x 10–3 – 3 x 10–4) (Boudsocq et al., 2001) and in vivo (Brotcorne-Lannoye and Maenhaut-Michel, 1986; Kim et al., 1997; Wagner and Nohmi, 2000). This property of Y-family DNA polymerase prompted us to evaluate its potential for preparing frame shuffling libraries.
Dpo4 is a Y-family of DNA polymerase isolated from the crenarchaeon Sulfolobus solfataricus P2 (Boudsocq et al., 2001) and is akin to human pol 
, inactivation of which leads to increased susceptibility to sunlight-induced skin cancer (Johnson et al., 1999; Masutani et al., 1999). Dpo4 is the first thermostable Y-family enzyme isolated and Boudsocq et al. have shown that this enzyme can substitute for Taq polymerase in polymerase chain reaction (PCR) (Boudsocq et al., 2001). We used this thermotolerant polymerase to prepare a mutant library, which we then compared with a library prepared using conventional error-prone PCR (EP-PCR) (Bartel and Szostak, 1993).
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionAcknowledgementsReferences |
|---|
Purification of Dpo4
Dpo4 was purified as described previously (Boudsocq et al., 2001) with slight modifications. We used Escherichia coli Rosetta(DE3) (Novagen, Madison, WI) as the host strain for the expression Dpo4 from p1914, which allowed efficient translation of genes containing rare codons. An overnight culture of Rosetta(DE3)/p1914 in Luria–Bertani medium (Sambrook and Russel, 2001) containing 50 µg/ml carbenicillin (Sigma Chemical, St Louis, MO) and 25 µg/ml chloramphenicol (Sigma Chemical) was diluted 1:100 in 100 ml of the same medium and incubated at 37°C. When the OD660 reached 0.4, isopropyl-ß-thio-D-galactopyranoside (final concentration, 1 mM; TaKaRa-Bio, Shiga, Japan) was added to the culture, which was then incubated for an additional 3 h. The cells were then harvested by centrifugation and stored at –80°C until used, at which time they were thawed, resuspended in buffer A {20 mM 2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES) pH 7.0, 50 mM NaCl, 0.1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM dithiothreitol (DTT)} and sonicated using a Biorupter (output 200 W, duty cycle = 2 min, 3–4 times; Cosmo Bio, Tokyo, Japan). After removing the cell debris by centrifugation, the cell extract was heated at 55°C for 10 min to precipitate any heat-labile proteins. The precipitate was removed by centrifugation (20 000 g for 10 min) and the supernatant was applied to a Hi-Load S 10/16 cation-exchange column (Amersham Pharmacia Biotech, Piscataway, NJ) equilibrated with buffer A. The proteins were fractionated using a linear gradient of 50–1000 mM NaCl in buffer A run at a flow rate of 5 ml/min in an FPLC system (Amersham Pharmacia Biotech). The enzyme eluted at 
0.5 M NaCl; 8 ml were collected, concentrated using a Centriprep 10 (Amicon, Beverly, MA) and stored at –20°C in 50% glycerol. The protein concentration was determined by measuring the OD280 using a UV2550 spectrophotometer (Shimadzu, Kyoto, Japan); the extinction coefficient was 19 200.
Frame shuffling PCR
PCR catalyzed by Dpo4 was carried out as described previously (Boudsocq et al., 2001) with slight modifications. First, linear DNA templates were prepared by standard PCR: the reaction mixture (50 µL) contained 300 pmol of each primer [KY-1087 (5'-GGA TAA CAA TTC CCC TCT AGA AAT-3') and KY-1086 (5'-TTG CTC AGC GGT GGC AGC AGC CAA-3')], 1.75 units of High-Fidelity Taq polymerase (Roche Diagnostics, Basel, Switzerland), 10 mM Tris–HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 50 µM dNTPs and template DNA (pYT320) that contained the repeated artificial protein gene. A thermal cycle comprising 94°C for 30 s, 55°C for 30 s and 72°C for 30 s was repeated 30 times in a GeneAmp 9600 temperature cycler (Perkin-Elmer, Norwalk, CT). The amplified linear DNAs were separated by TAE agarose gel electrophoresis and recovered from the gel using a GeneClean II kit (Qbiogene, Carlsbad, CA). The concentration of purified DNA was determined by ethidium bromide staining after re-electrophoresis in an agarose gel. The reaction mixture for frame shuffling PCR contained 10 ng of linear DNA template, 480 pmol of each primer (KY-1087 and KY-1086), 40 mM Tris–HCl (pH 8.0), 5 mM MgCl2 10 mM DTT, 60 mM KCl and 1 M betaine (Sigma Chemical). The mixture was first heated at 99.9°C for 5 min, then cooled to 90°C, after which 1.5 µg of Dpo4 (1 µl) was added to the mixture. This was followed by two cycles of 90°C for 30 s, 55°C for 20 s and 65°C for 3 min. The amplified DNAs were excised from the agarose gels and purified using a GeneClean II kit. Finally, standard PCR was again performed against the same linear template preparation using purified DNA as the template and a pair of nested inner primers [KY-837 (5'-AAT TTT GTT TAA CTT TAA GAA GGA GA-3') and KY-836 (5'-TCA GCT TCC TTT CGG GCT TTG TTA-3')]. The amplified DNA was digested with SalI and SpeI (New England Biolabs, Beverly, MA) and cloned into pKS600 (Saito et al., 2004), which was derived from the pQE9 vector series (Qiagen, Hilden, Germany). The DNA sequences were confirmed by sequencing about 20 clones from the resultant plasmid libraries.
Conventional error-prone PCR mutagenesis
Mismatched mutations were introduced using the PCR mutagenesis method of Bartel and Szostak (1993), which was a modification of the protocol of Cadwell and Joyce (1992). Briefly, DNA was amplified using Taq DNA polymerase in the presence of Mn2+ and the exponential growth of the DNA was sustained by serially diluting the amplified DNA with reaction mixture.
Purification and characterization of artificial proteins
Expression and purification of artificial proteins were carried out as described previously (Shiba et al., 2002b). Briefly, proteins were purified under denaturing conditions using Talon resin (Clontech, Palo Alto, CA) according to manufacturer's protocol. Eluted proteins were dialyzed against buffer containing 50 mM Tris–AcOH (pH 4.0), 100 mM NaCl and 1 mM EDTA and concentrated with a Centriprep 10.
CD spectra were recorded using a JASCO J-820 spectrophotometer (Japan Spectroscopic, Tokyo, Japan) with a cell having a 1 mm pathlength. Samples (5 µM) were prepared in buffer containing 50 mM sodium phosphate (pH 7.5). The circular dichroism (CD) spectra from 250 to 200 nm were recorded at 25°C and averaged over three accumulations.
Sedimentation equilibrium measurements were made using an XL-I analytical ultracentrifuge (Beckman, Fullerton, CA) equipped with a Ti60n rotor and a six-well centerpiece. Series of diluted proteins (20, 10 and 5 µM) were prepared in the same buffer as used for DLS measurements (see below). Samples were centrifuged first at 15 000 r.p.m. for 12 h at 20°C. By that time, the distribution of solute had reached equilibrium and radial scans were recorded at 230 nm. Thereafter, the rotation speed was set at 20 000 r.p.m. and then at 25 000 r.p.m. and the equilibrium was checked and radial scans recorded at each of those rotation speeds. The solvent density (1.035) and partial specific volume of each protein calculated from its amino acid composition based on the sequence in the absence of Gdm-HCl were used for the data analysis. To determine the average MW, recorded data obtained with three protein concentrations were globally fitted to a single ideal species using an analysis program developed by Beckman and installed on the analytical ultracentrifuge.
Dynamic light scattering (DLS) was carried out in a 12 µl sample cuvette using Dynapro MS-X or Dynapro 800 equipment (Protein Solutions, Piscataway, NJ). The samples were prepared in buffer containing 50 mM sodium phosphate (pH 7.5), 150 mM sodium chloride and 1 M Gdm.HCl. The solvent viscosity (1.043) and refractive index (1.3507) used were the reported values obtained with 1 M Gdm-HCl (Kawahara and Tanford, 1966). All measurements were performed at 20°C; data were recorded for 200 s.
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionAcknowledgementsReferences |
|---|
As a starting parent gene, we used an artificial gene, #320, which we created with our microgene-based method (Shiba et al., 2002b
). This 268 bp gene is composed of 6.5 tandem repeats of a microgene, MG-15, and has a single frame shift mutation in its 3' region, so that it gives a translational product with the structure (5.5xframe 1) – (1xframe 3) (Figure 1B). Because MG-15 has no termination codon in any of its reading frames, the untranslated frames of #320 also had no termination codons. After carrying out PCR with Dpo4 to amplify #320, progenitors were ligated into a vector and the sequences of 20 randomly selected clones were determined (Figure 2A). Among them, two had the wild-type sequences, while the remaining 18 contained 86 base substitution mutations (mutation rate = 1.6x10–2) and 24 base deletion mutations (4.7x10–3) (Table I). Most of the deletion mutations were the ‘one base deletion’ type, which is consistent with the previously reported properties of Y-family DNA polymerases (Boudsocq et al., 2001; Kokoska et al., 2002). Base substitution mutations resulted in 58 missense mutations, 23 silent mutations and five nonsense mutations among the translational products. Twelve of the clones exhibited frame shifts within their reading frames caused by ins/del mutations and, of those, six showed multiple frame shifts (the extreme example contained four frame shift mutations, FS-3 in Figure 2A).
|
|
For comparison, we also prepared a mutant library using conventional EP-PCR under moderate and highly mutagenic conditions. As summarized in Table I, the moderate conditions led to base substitutions at a rate of 2.2 x 10–2, but gave ins/del mutations at a rate of 8.7 x 10–4. Under highly mutagenic conditions, the ins/del mutation rate rose to 3.7 x 10–3, which was comparable to that obtained with PCR catalyzed by Dpo4, but the base substitution rate jumped to 7.1 x 10–2, leading to the appearance of fatal termination codons (Figure 2B). We were therefore better able to obtain a sufficient frame shift mutation rate while, at the same time, suppressing nonsense mutations with PCR catalyzed by a Y-family DNA polymerase than with conventional error-prone PCR. We named PCR mutagenesis mediated by a Y-family DNA polymerase ‘frame shuffling’ PCR. Under the conditions used, translation of seven of the 20 clones terminated prematurely because of the appearance of stop codons. On the other hand, some types of base modification (e.g. tetrahydrofuran) (Kokoska et al., 2003
) increased the frame shift mutation rate, thereby suppressing somewhat the appearance of nonsense mutations and improving the quality of the library.
To analyze the physicochemical properties of the mutant proteins obtained by frame-shuffling PCR, we chose four mutants that contained frame shift mutations (FS-8, FS-10, FS-11 and FS-13) and two that contained only missense mutations (MS-5 and MS-18) (Figures 1B and 2A) and investigated the secondary structures, molecular weights (MW) and hydrodynamic radii of their translational products using CD spectroscopy, sedimentation equilibrium centrifugation and DLS, respectively (Figure 3 and Table III). All of these mutant proteins were soluble in phosphate buffer of pH 4.0, 6.0 or 7.5 at 5 µM protein concentrations. As summarized in Table II, the proteins produced from parental #320 had an 
-helix-rich structure and an average MW of 66 760 (Figure 3 and Table III). That #320 encodes a polypeptide with an MW of 12 041 means that the protein forms oligomers composed, on average, of 5.3 polypeptides. This oligomerization state was consistent with the hydrodynamic radius observed with DLS (Table II) and the Stokes radius (44.8 Å) obtained from size-exclusion chromatography (data not shown). The -helical and oligomeric structures were retained by both MS-5 and MS-18, whereas the frame shuffled mutants showed distinct new properties taken on as the proteins acquired segments from other frames. For instance, FS-10 and FS-11 showed less -helicity and FS-13 showed a CD spectrum typical of a disordered peptide. Some of the mutants that contained peptide sequences from the frame 2 have been predicted to have high helical contents (i.e. FS-8 and FS-11). However, the observed helical contents in CD were much less than those calculated. The differences between observed and calculated helical contents showed the difficulty in predicting secondary structure from artificial sequences.
|
|
|
The oligomerization state also changed in the mutants, as FS-10 and FS-13 formed dimers under the test conditions. Thus, segmental substitution led to significant alterations in the physicochemical properties of the proteins.
In this work, we developed a novel method for preparing artificial protein libraries in which a Y-family DNA polymerase is used to introduce frame shift mutations on a parental gene. If the parental gene is devoid of termination codons in its unused reading frames, the resultant progeny genes have long open reading frames composed of scrambled sequences of its parental three reading frames. The libraries could provide novel types of mutant proteins that it may be difficult to obtain with a conventional base substitution mutagenesis.
Recently, microgene-based mutagenesis or exon-shuffling type mutagenesis has attracted attention for preparing artificial protein libraries (Kolkman and Stemmer, 2001; Lutz et al., 2001; Kitamura et al., 2002; Shiba et al., 2002a; Bittker et al., 2004; Tsuji et al., 2004), in which blocks of short DNA sequences are combinatorially polymerized to make long, open reading frames. In these experiments, unwanted introduction of frame shift mutation, which is inherent in DNA recombination experiments, could lower the quality of the libraries (Davidson et al., 1995; Prijambada et al., 1996; Cho et al., 2000). However, if the DNA blocks are prepared so that none of their reading frames would contain termination codon, we can easily prepare libraries that eliminate termination codons in all of their reading frames. Furthermore, if the DNA blocks were defined so that two or three of their frames would contain certain biological or structural information (Shiba, 2004), frame shuffling may provide opportunities for exploring larger areas of sequence space in exon shuffling-type experiments.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionAcknowledgementsReferences |
|---|
We thank Drs F.Boudsocq and R.Woodgate kindly for providing the Dpo4 expression plasmid and Drs H.Ohmori and K.Sano for helpful discussions. Funding to pay the Open Access publication charges for this article was provided by CREST, JST.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionAcknowledgementsReferences |
|---|
Bartel,D.P. and Szostak,J.W. (1993) Science, 261, 1411–1418.
Bittker,J.A., Le,B.V., Liu,J.M. and Liu,D.R. (2004) Proc. Natl Acad. Sci. USA, 101, 7011–7016.
Boudsocq,F., Iwai,S., Hanaoka,F. and Woodgate,R. (2001) Nucleic Acids Res., 29, 4607–4616.
Boudsocq,F., Kokoska,R.J., Plosky,B.S., Vaisman,A., Ling,H., Kunkel,T.A., Yang,W. and Woodgate,R. (2004) J. Biol. Chem., 279, 32932–32940.
Brotcorne-Lannoye,A. and Maenhaut-Michel,G. (1986) Proc. Natl Acad. Sci. USA, 83, 3904–3908.
Cadwell,R.C. and Joyce,G.F. (1992) PCR Methods Appl., 2, 28–33.[Medline]
Cho,G., Keefe,A.D., Liu,R., Wilson,D.S. and Szostak,J.W. (2000) J. Mol. Biol., 297, 309–319.[CrossRef][ISI][Medline]
Cirino,P.C., Mayer,K.M. and Umeno,D. (2003) Methods Mol. Biol., 231, 3–9.[Medline]
Davidson,A.R., Lumb K.J. and Sauer R.T. (1995) Nat. Struct. Biol., 2, 856–864.[CrossRef][ISI][Medline]
Friedberg,E.C., Wagner,R. and Radman,M. (2002) Science, 296, 1627–1630.
Goodman,M.F. and Tippin,B. (2000) Nat. Rev. Mol. Cell Biol., 1, 101–109.[CrossRef][ISI][Medline]
Johnson,R.E., Kondratick,C.M., Prakash,S. and Prakash,L. (1999) Science, 285, 263–265.
Kashiwagi,K. and Shiba,K. (2004) J. Mol. Catal. B, 28, 215–221.[CrossRef]
Kawahara,K. and Tanford,C. (1966) J. Biol. Chem., 241, 3228–3232.
Kim,S.R., Maenhaut-Michel,G., Yamada,M., Yamamoto,Y., Matsui,K., Sofuni,T., Nohmi,T. and Ohmori,H. (1997) Proc. Natl Acad. Sci. USA, 94, 13792–13797.
Kitamura,K., Kinoshita,Y., Narasaki,S., Nemoto,N., Husimi,Y. and Nishigaki,K. (2002) Protein Eng., 15, 843–853.
Kokoska,R.J., Bebenek,K., Boudsocq,F., Woodgate,R. and Kunkel,T.A. (2002) J. Biol. Chem., 277, 19633–19638.
Kokoska,R.J., McCulloch,S.D. and Kunkel,T.A. (2003) J. Biol. Chem., 278, 50537–50545.
Kolkman,J.A. and Stemmer,W.P. (2001) Nat. Biotechnol., 19, 423–428.[CrossRef][ISI][Medline]
Lutz,S., Ostermeier,M., Moore,G.L., Maranas,C.D. and Benkovic,S.J. (2001) Proc. Natl Acad. Sci. USA, 98, 11248–11253.
Masutani,C., Kusumoto,R., Yamada,A., Dohmae,N., Yokoi,M., Yuasa,M., Araki,M., Iwai,S., Takio,K. and Hanaoka,F. (1999) Nature, 399, 700–704.[CrossRef][Medline]
Neylon,C. (2004) Nucleic Acids Res., 32, 1448–1459.
Ohashi,E., Ogi,T., Kusumoto,R., Iwai,S., Masutani,C., Hanaoka,F. and Ohmori,H. (2000) Genes Dev., 14, 1589–1594.
Ohmori,H., et al. (2001) Mol. Cell, 8, 7–8.[CrossRef][ISI][Medline]
Prijambada,I.D., Yomo,T., Tanaka,F., Kawama,T., Yamamoto,K., Hasegawa,A., Shima,Y., Negoro,S. and Urabe,I. (1996) FEBS Lett., 382, 21–25.[CrossRef][ISI][Medline]
Rost,B. and Sander,C. (1993) J. Mol. Biol., 232, 584–599.[CrossRef][ISI][Medline]
Saito,H., Honma,T., Minamisawa,T., Yamazaki,K., Noda,T., Yamori,T. and Shiba,K. (2004) Chem. Biol., 11, 765–773.[CrossRef][ISI][Medline]
Sambrook,J. and Russel,D.W. (eds) (2001) Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Shiba,K., Takahashi,Y. and Noda,T. (1997) Proc. Natl Acad. Sci. USA, 94, 3805–3810.
Shiba,K., Hatada,T., Takahashi,Y. and Noda,T. (2002a) J. Biochem. (Tokyo), 132, 689–696.
Shiba,K., Takahashi,Y. and Noda,T. (2002b) J. Mol. Biol., 320, 833–840.[CrossRef][ISI][Medline]
Shiba,K. (2004) J. Mol. Catal. B, 28, 145–153.[CrossRef]
Strauss,B.S., Roberts,R., Francis,L. and Pouryazdanparast,P. (2000) J. Bacteriol., 182, 6742–6750.
Tsuji,T., Onimaru,M., Kitagawa,M., Kojoh,K., Tabata,N. and Yanagawa,H. (2004) Methods Enzymol., 388, 61–75.[ISI][Medline]
Wagner,J., Gruz,P., Kim,S.R., Yamada,M., Matsui,K., Fuchs,R.P. and Nohmi,T. (1999) Mol. Cell, 4, 281–286.[CrossRef][ISI][Medline]
Wagner,J. and Nohmi,T. (2000) J. Bacteriol., 182, 4587–4595.
Zhang,Y., Yuan,F., Wu,X., Wang,M., Rechkoblit,O., Taylor,J.S., Geacintov,N.E. and Wang,Z. (2000) Nucleic Acids Res., 28, 4138–4146.
Received October 5, 2005; revised December 7, 2005; accepted December 8, 2005.
Edited by Taiji Imoto
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Saito, S. Kashida, T. Inoue, and K. Shiba The role of peptide motifs in the evolution of a protein network Nucleic Acids Res., October 8, 2007; 35(19): 6357 - 6366. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||