Skip Navigation



PEDS Advance Access published online on February 20, 2008
Protein Engineering Design and Selection, doi:10.1093/protein/gzn004
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
21/4/267    most recent
gzn004v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Emond, S.
Right arrow Articles by Remaud-Simeon, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Emond, S.
Right arrow Articles by Remaud-Simeon, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

A novel random mutagenesis approach using human mutagenic DNA polymerases to generate enzyme variant libraries

Stéphane Emond1,2,3, Philippe Mondon4, Sandra Pizzut-Serin1,2,3, Laurent Douchy4, Fabien Crozet4, Khalil Bouayadi4, Hakim Kharrat4, Gabrielle Potocki-Véronèse1,2,3, Pierre Monsan1,2,3 and Magali Remaud-Simeon1,2,3,5

1 Université de Toulouse; INSA, UPS, INP; LISBP, 135 Avenue de Rangueil, F-31077 Toulouse, France 2INRA, UMR792 Ingénierie des Systèmes Biologiques et des Procédés, F-31400 Toulouse, France 3CNRS, UMR5504, F-31400 Toulouse, France 4MilleGen SA, Immeuble BIOSTEP, Bâtiment A, rue Pierre et Marie Curie, BP 38183, 31681 Labège cedex, France

5 To whom correspondence should be addressed. E-mail: magali.remaud{at}insa-toulouse.fr


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results and discussion
 Conclusion and future...
 Funding
 Acknowledgements
 References
 
The in vitro MutaGenTM procedure is a new random mutagenesis method based on the use of low-fidelity DNA polymerases. In the present study, this technique was applied on a 2 kb gene encoding amylosucrase, an attractive enzyme for the industrial synthesis of amylose-like polymers. Mutations were first introduced during a single replicating step performed by mutagenic polymerases pol β and pol {eta}. Three large libraries (>105 independent clones) were generated (one with pol β and two with pol {eta}). The sequence analysis of randomly chosen clones confirmed the potential of this strategy for the generation of diversity. Variants generated by pol β were 4–7-fold less mutated than those created with pol {eta}, indicating that our approach enables mutation rate control following the DNA polymerase employed for mutagenesis. Moreover, pol β and pol {eta} provide different and complementary mutation spectra, allowing a wider sequence space exploration than error-prone PCR protocols employing Taq polymerase. Interestingly, some of the variants generated by pol {eta} displayed unusual modifications, including combinations of base substitutions and codon deletions which are rarely generated using other methods. By taking advantage of the mutation bias of naturally highly error-prone DNA polymerases, MutaGenTM thus appears as a very useful tool for gene and protein randomisation.

Keywords: amylosucrase/directed evolution/low-fidelity DNA polymerase/protein randomisation/random mutagenesis


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results and discussion
 Conclusion and future...
 Funding
 Acknowledgements
 References
 
Over the last two decades, directed evolution has become a powerful approach for tailoring protein features to industrial needs and probing structure–activity relationships (Yuan et al., 2005Go). Inspired by the Darwinian concept of natural evolution, this methodology consists in constructing libraries of variant genes followed by screening and/or selection of their protein products. In vitro creation of libraries enables the exploration of sequence space independently from a selective pressure under physiological conditions and thus allows the search for variants adapted to non-natural and harsh environment close to those found in industrial processes. One of the most effective strategies for library construction is to randomly mutate the gene encoding the parental protein. This approach requires no structural or mechanistic information and can uncover unexpected beneficial mutations. Several in vitro or in vivo random mutagenesis methods have been described so far (Wong et al., 2006Go). Among these, error-prone PCR (epPCR) techniques are the most popular due to their simplicity (Leung et al., 1989Go; Cadwell and Joyce, 1992Go; Zaccolo et al., 1996Go; Lin-Goerke et al., 1997Go; Shafikhani et al., 1997Go; Cline and Hogrefe, 2000Go).

DNA polymerases catalyse the template-directed incorporation of deoxynucleotides into a growing primer terminus to generate a new complementary daughter strand. They have been classified on the basis of their structural similarities into seven families. Among them, family A, B, C, D and Y share a common structural organisation that can be described as a right hand with three distinct domains referred to as ‘palm’, ‘fingers’ and ‘thumb’ (Beard and Wilson, 2003Go). The active site is located within the palm domain. The fingers and thumb domains are important for dNTP recognition/binding and correct positioning of the DNA substrate, respectively. Crystallographic studies of Y family DNA polymerases (Ling et al., 2001Go; Silvian et al., 2001Go; Trincao et al., 2001Go; Zhou et al., 2001Go) have revealed that these enzymes differ from the other polymerases in having rather small fingers and thumb domains, which lead to an open and solvent accessible active site in the palm domain. Moreover, Y family polymerases possess an additional domain referred to as ‘little finger’, which determines the catalytic efficiency and mutation spectra of each Y family member by influencing enzyme–substrate interactions (Boudsocq et al., 2004Go). Beyond this common structural framework, DNA polymerases from family X are described as ‘left-handed’ because their palm domain is not homologous with those of the other polymerase families (Beard and Wilson, 2003Go).

The fidelity of DNA polymerases, defined as the inverse of misinsertion frequency, is ensured by base pair complementarity, by substrate-induced conformational changes and, in some cases, by proofreading catalysed by a 3’->5’ exonuclease domain (Kunkel, 2004Go). Such fidelities span a wide range. Replicative DNA polymerases from structural families A, B and C are high fidelity enzymes exhibiting error frequencies of ~10–6 (one error per million nucleotides incorporated). The most commonly used polymerase in epPCR methods, Taq DNA polymerase (Taq pol) from family A, displays an error rate that ranges between 10–4 and 10–5 (Tindall and Kunkel, 1988Go; Eckert and Kunkel, 1990Go). The least accurate DNA polymerases belong to the X and Y families (Prakash et al., 2005Go; Moon et al., 2007Go). These enzymes are involved in DNA damage repair mechanisms. Mammalian DNA polymerase β (pol β), which is implicated in the base excision repair (BER) pathway (Matsumoto and Kim, 1995Go), is the best known member of the X-family and exhibits an error frequency ranging from 10–3 to 10–4 (Kunkel, 1985Go; Osheroff et al., 1999Go). Y-family polymerases, also known as translesion synthesis (TLS) polymerases, replicate DNA in a distributive manner and lack any exonucleolytic activity for proofreading (Kunkel, 2004Go). These enzymes display the highest error rates measured among DNA polymerases (10–1–10–3) (Kunkel, 2004Go). In human, known members of this family are DNA polymerases {eta} (pol {eta}), {iota} or {kappa} (Kunkel et al., 2003Go).

The hypermutation profile of these repair DNA polymerases has been used to design a new in vitro random mutagenesis technique, MutaGenTM (Bouayadi et al., 2002Go). This method consists of a first single mutagenic replication step catalysed by a low-fidelity human DNA polymerase, which is followed by a selective PCR amplification of the replicated mutated sequences (Fig. 1). MutaGenTM was recently reported as an efficient method to enlarge the initial diversity of human fragment antibody libraries (Mondon et al., 2007Go). It has, however, never been used on templates longer than 700 bp. Here we report the use of the MutaGenTM process either with human pol β or pol {eta} to produce random variants of amylosucrase from Neisseria polysaccharea (AS). This enzyme encoded by a 2 kb gene is an attractive biocatalyst for the in vitro synthesis of amylose-like polymers using sucrose as sole substrate (Potocki de Montalk et al., 1999Go). Three large amylosucrase variant libraries were constructed using MutaGenTM. The generated diversity was analysed from sequencing results of clones randomly chosen in the libraries. The advantages of this novel approach are discussed on the basis of the provided results.


Figure 1
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. The MutaGenTM process. MutaGenTM consists of a first single mutagenic replication step followed by selective PCR amplification of the replication products. This technique uses primers possessing a 5' extension (coloured in red) that is not complementary to the template. After template denaturation and primers hybridisation, replication is carried out by a low-fidelity DNA polymerase. The mutated copies (coloured in green) are then selectively amplified by a PCR procedure consisting of a first cycle with a low hybridisation temperature followed by selective cycles (up to 25) with high hybridisation temperature disabling template amplification. This methodology was applied to construct AS variant libraries B and C using pol {eta}.

 

   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results and discussion
 Conclusion and future...
 Funding
 Acknowledgements
 References
 
Bacterial strains and plasmids

One Shot® Escherichia coli TOP10 (Invitrogen) was used for transformation of ligation mixtures. Escherichia coli JM109 (Promega) was used to determine the fraction of active AS clones present in the random mutagenesis libraries. Plasmid pGEX-6P3 (Amersham Pharmacia Biotech) was used for the cloning of libraries A and B. Plasmid pGST-AS, derived from pGEX-6P-3 (Potocki de Montalk et al., 1999Go) was used as site-directed mutagenesis template to introduce the restriction sites necessary for the cloning of library C. The AS encoding sequence inserted in plasmid pCEASE01 (van der Veen et al., 2004Go) was used as replication template for libraries A and B.

DNA manipulations

Restriction endonucleases and DNA-modifying enzymes were purchased from New England Biolabs and used according to the manufacturer’s instructions. DNA purification was performed using QIAQuick (PCR purification and gel extraction) and QIASpin (miniprep and maxiprep) (Qiagen). DNA sequencing was carried out using the dideoxy chain termination procedure by MilleGen sequencing service (Labège, France).

Site-directed mutagenesis

Site-directed mutagenesis was carried out with the QuikChangeTM site-directed mutagenesis kit (Stratagene) as previously described (Sarcabal et al., 2000Go) to introduce silent mutations creating restriction sites HindIII and XhoI into the AS encoding sequence. Plasmid pGST-AS was used as template. The pairs of primers used for each mutagenesis inverted PCR were (restriction sites are underlined): AS-XhoI-for, 5'-GCGGCTTCTCGCAACTCGAGGACGGACGCTGGGTGT-3'; AS-XhoI-rev, 5'-ACACCCAGCGTCCGTCCTCGAGTTGCGAGAAGCCGC-3'; AS-HindIII-for, 5'-GCCGTTGACCGCATCAAGCTTTTGTACAGCATTGCTTTG-3'; AS-HindIII-rev, 5'-CAAAGCAATGCTGTACAAAAGCTTGATGCGGTCAACGGC-3'. The resulting plasmid pGST-AS(X-H) was used as replication template for library C and for cloning the mutated replication products.

Construction of AS libraries using the MutaGenTM procedure

A mixture containing 1 µg of template plasmid, primers in equimolar concentration (200 nM) and the appropriate replication buffer was treated for 5 min at 95°C and immediately cooled down at 4°C to denature DNA strands. Following this step, mutagenic replications were started by adding 50 µM dATP/dCTP, 100 µM dTTP/dGTP and 4 units of human pol β (for library A) or 100 µM dNTPs and 4 units of pol {eta} (for libraries B and C). One unit is defined as the amount of enzyme required to catalyse the incorporation of 1 nmol of dNTP into an acid-insoluble form in 1 h at 37°C. These reactions were performed for 2 h at 37°C. Four primers were used for library A (italic characters indicate a tail that is not complementary with the template): AmS-F, 5'-ACGATGCCTGCAGGTCGTGCAAGCTTGTAATGAATTCACAGTA-3' (EcoRI site underlined); AmS-R, 5'-GAGGGCACATGGATCAGACCACTCTAGATGCATGCTCGAGAA-3' (XhoI site underlined); ASint-F, 5'-TACAGCAACCCGTGGGTATT-3'; ASint-R, 5'-AATACCCACGGGTTGCTGTA-3'. The primers used for library B were AmS-F and AmS-R. Two primers were used for library C to replicate a region comprised between nucleotides 726 and 1410 of the as gene: ASk7-F, 5'-TGCTCGTGATCGTGACGCTATAACTCGAGGACGGACGCT-3' (XhoI site underlined); ASk7-R, 5'- AGCTGCTGCACGCTCGTGACTACAAAAGCTTGATGCGGTCAA-3' (HindIII site underlined). Composition of pol β replication buffer composition was 50 mM Tris–Cl pH 8.8, 10 mM MgCl2, 10 mM KCl, 1 mM DTT, 1% (v/v) glycerol. Replication buffer of pol {eta} was 25 mM Tris HCl pH 7.2, 1 mM DTT, 5 mM MgCl2, 2.5% (v/v) Glycerol. Human DNA polymerases β and {eta} were produced and purified as described in the supplementary material available at PEDS online.

After the first step, two strategies were employed to amplify the mutated replication products (Figs 1 and 2). For library A, two groups of fragments generated by pol β (corresponding to two separate parts of AS encoding sequence) were amplified with Platinum Taq polymerase (Invitrogen) in two different PCRs, one with primer pair AmsF/ASintR and the other with ASintF/AmsR (Fig. 2). Owing to their terminal complementarities, the two groups of amplified replication products were extensively amplified with Platinum Taq in a first PCR without primers (25 cycles of 20 s at 94°C, 10 s at 57°C and 2 min at 72°C) followed by a second PCR with primers AmS-F and AmS-R (25 cycles of 20 s at 94°C and 1 min 30 s at 72°C).


Figure 2
View larger version (20K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. Modified MutaGenTM methodology to construct full-length as gene library using pol β. After template denaturation and primers hybridisation, two parts of the template DNA bearing an overlapping region are replicated by pol β in separate reactions. After PCR amplification, the mutated copies (coloured in green) are finally reassembled into a full-length gene during an overlap-extension PCR step. This methodology was applied to construct AS variant library A using pol β.

 
The fragments generated by pol {eta} were amplified using Platinum Taq with primers AmS-F/AmS-R (library B) or ASk7-F/ASk7-R (library C) through a single selective PCR process (Fig. 1; first cycle: 2 min at 94°C, 10 s at 58°C and 2 min at 72°C; then 25 cycles: 20 s at 94°C and 1 min 30 s at 72°C).

The amplified replication products were cloned into EcoRI and XhoI restriction sites of pGEX-6P3 (libraries A and B) or into XhoI and HindIII restriction sites of pGST-AS(X-H) (library C). The ligation mixtures were transformed into electrocompetent E.coli TOP10 cells subsequently plated on solid LB medium with 100 µg/ml ampicillin. After growth, the number of colonies was determined to estimate the size of the libraries and some were randomly chosen for DNA sequencing. For library storage, the colonies were scraped with sterilised glycerol 10% (v/v) and conserved in this form at –80°C. Alternatively, the colonies were used for plasmid extraction to store the libraries in the form of DNA solution.

Construction of an AS variant library by epPCR

epPCR using Taq DNA polymerase (New England Biolabs) was performed using wild-type AS encoding sequence as template and the primer pair AmsF/AmsR. The reaction was carried out from 5 ng of full pCEASE01vector with 5 U of Taq pol in 1x ThermoPol buffer (New England Biolabs), 7 mM MgCl2, 200 nM of each primer, 0.2 mM dATP/dGTP and 1 mM dCTP/dTTP. The products were cloned into EcoRI and XhoI restriction sites of pGEX-6P3. Ligation mixtures were transformed into electrocompetent E.coli TOP10 cells subsequently plated on solid LB medium with 100 µg/ml ampicillin.

Determination of amylose-producing clone fraction in a library of variants

AS variant libraries were transformed into E.coli JM109 cells. The colonies were plated and grown on solid LB medium containing 50 g/l sucrose, 1 mM IPTG, 100 µg/ml ampicillin. The colonies expressing active AS variants were detected by exposing the plates to iodine vapours, allowing an estimation of the amount of amylose-producing clones present in each library.

Analysis of sequencing data

The sequences of clones randomly chosen in each library were analysed using MilleGen proprietary software Mutanalyse 2.5. To characterise library diversity, various parameters were determined from the alignment of the sequenced clones both at the nucleotide and the protein levels. The mutation, base substitution, deletion and addition rates were calculated as the ratio between the number of corresponding events and the number of bases sequenced. The mutational spectrum consisted in listing the different mutation subtypes (additions, deletions and substitutions). Mutanalyse 2.5 also determined the impact of the mutations at the protein level. Clones from library A were sequenced over the entire encoding sequence. Clones from library B were sequenced on the first 552 bp. Clones from library C were sequenced over the region targeted for random mutagenesis (between nucleotides 726 and 1410). Variants generated by epPCR were sequenced over the entire encoding sequence.


   Results and discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
Conclusion and future...
 Funding
 Acknowledgements
 References
 
Construction of AS variant libraries using pol β and pol {eta}

The MutaGenTM technique was applied to generate three libraries (named A, B and C) using wild-type as gene as parental template. Libraries A and B are full-length gene libraries constructed using pol β and pol {eta}, respectively. The initial procedure illustrated in Fig. 1 was modified to overcome pol β low processivity (Osheroff et al., 1999Go). As shown in Fig. 2, two parts of the as gene bearing an overlapping region were first replicated by pol β during separate reactions. This step was performed in the presence of unbalanced dNTPs concentrations to increase the polymerase in vitro error rate (Sinha and Haimes, 1981Go). The replication products were then amplified by PCR and reassembled into a full-length gene during a final overlap-extension PCR step (Fig. 2). Conversely, the higher processivity of pol {eta} (Bebenek et al., 2001Go) allowed replicating the full-length as sequence during a single reaction to generate library B (Fig. 1). Library C was constructed using pol {eta} to focus the introduction of mutations into a region of AS catalytic domain (from residues 242 to 470) that includes many critical residues for catalytic activity and substrate specificity (Sarcabal et al., 2000Go; Skov et al., 2002Go; Albenne et al., 2004Go). The cloning of the amplified replication products into the pGEX vector resulted in libraries A, B and C of 8 x 105, 1 x 105 and 1 x 105 independent clones, respectively. Prior to any selection, several clones from each library were randomly chosen and sequenced in order to analyse the diversity generated by the MutaGenTM approach (Table I).


View this table:
[in this window]
[in a new window]

  		Table I. Sequence analysis and mutation rates of AS variant libraries generated by MutaGenTM (libraries A, B and C) or by epPCR

 
Mutation rates in AS variant libraries created with pol β and pol {eta}

As shown in Table I, the mutation rate measured for library A (1.7 x 10–3) was 4–7-fold lower than that of libraries B and C (11.2 x 10–3 and 7.8 x 10–3, respectively). These values indicate that the mutational charge (i.e. the number of mutations per variant) generated with pol {eta} was higher than that generated by pol β. This is in accordance with the reported pol {eta} error rate which is 10–100-fold higher than that of pol β (Matsuda et al., 2001Go). Interestingly, these mutagenic DNA polymerases allow the generation of diversity through a single replication step. For comparison, an AS variant library was created by epPCR using Taq pol in the presence of biased dNTPs concentrations (Table I). Thirty amplification cycles were needed to reach a mutation rate of 6.3 x 10–3, similar to those measured for the two libraries generated with pol {eta}. The addition of Mn2+ or the use of various unbalanced dNTPs concentrations are known to increase Taq pol error rate but these modifications generally reduce the yield of the epPCR reaction (Cadwell and Joyce, 1992Go; Vartanian et al., 1996Go; Lin-Goerke et al., 1997Go; Shafikhani et al., 1997Go). Thus, employing error-prone DNA polymerases like pol β and pol {eta} is clearly an advantage compared with epPCR approaches which are based on serial replication steps to increase the mutation rate.

Base substitution spectra

The mutations generated by both pol β and pol {eta} were predominantly single base substitutions (95.5%, 91.4% and 88.4% of the total number of mutations in libraries A, B and C, respectively). The ratio of transitions (Ts; base substitutions: purine to purine or pyrimidine to pyrimidine) to transversions (Tv; base substitutions: purine to pyrimidine or pyrimidine to purine) is a commonly used indicator to analyse the bias affecting the different types of base substitutions produced by a random mutagenesis method. For instance, a mutational spectrum with equiprobable base substitutions will exhibit a Ts/Tv ratio equal to 0.5. The rates of all kinds of base substitutions were determined from the clones sequenced in the present study: transitions represented 67.1–81.7% of mutations and transversions 18.3–32.9%. The values of the Ts/Tv ratio for libraries A, B and C were 3.7, 4.5 and 2, respectively. For epPCR using Taq pol, this ratio was shown to vary from 1.1 to 3.8 according to the reaction conditions listed in Table II. The higher rate of transitions compared with that of transversions is a common feature among random mutagenesis techniques that use DNA polymerases (Wong et al., 2006Go). Indeed, purine:pyrimidine and pyridimidine:purine mispairs induce less geometric distortions in DNA than purine:purine and pyrimidine:pyrimidine mispairs during nucleotide incorporation by DNA polymerases.


View this table:
[in this window]
[in a new window]

  		Table II. Comparison of base substitution spectra of MutaGenTM and epPCR methods

 
As shown in Table II, transitions were generated more homogeneously by pol β AT->GC: 43.9%; GC->AT: 34.8%) in comparison with pol {eta} which shows a bias towards AT->GC substitutions. The most frequent transversions produced by pol β were GC->TA (9%), AT->TA (6.6%) and AT->CG (4.2%), whereas GC->CG mutations occurred with a rate of 1.4%. Using pol {eta}, GC->TA (1.8% and 4.9% in libraries B and C, respectively) was the least observed transversion, whereas the most frequently generated were AT->TA (6.5% and 13.1%), GC->CG (3% and 8.1%) and AT->CG (2.3% and 6.5%). In comparison, the use of various epPCR conditions (Table II) does not modify the tendencies of Taq pol mutational spectrum, which is biased towards AT->GC transitions (from 35% to 65.2%) and AT->TA transversions (from 11.4% to 37.4%). The Mutazyme DNA polymerase which is a low-fidelity mutant of Taq pol is less biased for transitions (AT->GC: 22.2%; GC->AT: 27.8%). This enzyme also displays the same bias as Taq pol for AT->TA transversions (13.9%) but alternatively generates more frequently GC->TA substitutions (19.4%). The MutaGenTM approach thus provides different and complementary base substitution spectra and takes advantage of the different catalytic properties of pol β and pol {eta} which are probably related to their specific structural features.

Insertion and deletion mutations

Both pol β and pol {eta} generated insertions and deletions (indels) of bases (Table I). A 20-fold higher deletion rate was observed for variants created with pol {eta} (~6.7 x 10–4) in comparison with those obtained with pol β (0.3 x 10–4). Similarly, the insertion rate of pol {eta} was 5–7-fold higher (3.0 x 10–4 and 2.2 x 10–4 in libraries B and C, respectively) compared with pol β (0.4 x 10–4). The indel events observed were mostly single-base indels (Table I). Interestingly, some variants produced by pol {eta} were affected by deletions of two or more consecutive bases whereas all the deletions listed in the mutants generated by pol β were single-base deletions (Table I and see Table S1 in the Supplementary material available at PEDS online).

Most of the indel mutations introduced by pol β (8 out of 10 observed; see Table S1 available at PEDS online) were deletion or insertion of repeated nucleotides. Conversely, pol {eta} generated a higher number of this type of mutation in non-repetitive sequences (13 out of 24 observed; Table S1 available at PEDS online). Some of our results are thus in agreement with the classical explanation that frameshift mutations result from strand slippage during synthesis in repetitive DNA sequences (Streisinger et al., 1966Go). Alternative mechanisms explaining indels of either repeated or non-repeated nucleotides have also been proposed (Kunkel, 2004Go), including misalignment in the active site which is supported by structural evidence for Sulfolobus solfataricus Dpo4, a family Y DNA polymerase (Ling et al., 2001Go). However, the mechanisms by which complex indels involving two or more nucleotides arise are still unknown (Garcia-Diaz and Kunkel, 2006Go).

Mutational impact at the protein level

The sequenced AS variants were analysed at the protein level to further evaluate the diversity introduced by our method. Because of the high frameshift mutation rate of pol {eta}, sequences containing deletions and/or additions were more numerous in libraries B and C (40% and 30.8%, respectively; Table III) than in library A (11.3%; Table III). Since clones from library B were sequenced on a portion of 543 bp length (181 residues), this library should contain more than 40% of AS variants affected by frameshift mutations. Table IV summarises the effect of base substitutions observed in the sequenced variants at the protein level. Silent substitutions represented 35.1%, 36.5% and 16.4% in libraries A, B and C, respectively (Table IV). Owing to the structure of the genetic code, most of these silent mutations were transitions occurring on codon third positions (see Table S2 in the Supplementary material available at PEDS online). These events lowered the effective amino acid substitution rates which were 0.3%, 1.75% and 1.4% in libraries A, B and C, respectively (%: per 100 residues; Table IV). Remarkably, most transversions (88.2%; see Table S2 available at PEDS online) lead to amino acid mutations. Among amino acid substitutions, stop codons (nonsense substitution) were generated with relatively weak frequencies (Table IV; library A: 3.3%; library B: 4.7%; library C: none observed).


View this table:
[in this window]
[in a new window]

  		Table III. Protein sequence analysis of clones from AS libraries generated by MutaGenTM

 

View this table:
[in this window]
[in a new window]

  		Table IV. Impact of base substitutions in the sequenced AS variants at the protein level

 
The amount of variants retaining AS activity in each library was determined by detecting AS variants expressing colonies able to produce amylose when grown on solid-rich medium in the presence of sucrose. These rates were 75%, 20% and 35% for libraries A, B and C, respectively. These values can be related to the mutation rates applied in the libraries. Indeed, library A displayed the lowest mutation rate (1.7 x 10–3) and contained the highest amount of clones retaining AS activity. Conversely, active AS clones were less numerous in libraries B and C, for which higher mutation rates were measured. In addition, because of its high frameshift mutation rate, pol {eta} generates statistically more deleterious mutations than pol β. Indeed, most of the indel events observed were single-base indels which generally lead to protein truncation and inactivation. On the other hand, the ability of pol {eta} to generate deletions of two or more consecutive bases also led to variants affected by deletions of amino acid residues (see clones 24B and 8C in Fig. 3) or modifications of larger fragments (see clone 12C in Fig. 3B). Insertion and deletion of amino acids is an important mutational mechanism in protein evolution (Grishin, 2001Go; Taylor et al., 2004Go). Such mutations can have considerable effects on the properties of a protein, especially recognition, activity, specificity and stability. Recent examples of proteins that have been engineered through deletion/insertion of codons include antibodies (Lantto and Ohlin, 2002Go), TEM-1 β-lactamase (Simm et al., 2007Go) or Green Fluorescent Protein (Flores-Ramirez et al., 2007Go). A number of experimental strategies have been described to randomly insert or delete codons into a target protein (Murakami et al., 2002Go; Osuna et al., 2004Go; Jones, 2005Go; Fujii et al., 2006Go). The use of pol {eta} provides a simple approach to generate variants combining base substitutions with insertion and/or deletions (Fig. 3). The MutaGenTM method thus enables a wide sequence space exploration by producing mutated proteins displaying unusual modifications which are very rarely generated with other methods.


Figure 3
View larger version (54K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3. Distribution of amino acid mutations in AS variants generated with pol {eta}. Sequenced variants from libraries B and C are aligned with wild-type AS (AS WT). For the sake of clarity, only variants displaying in-frame mutations are represented. Primary structure elements (N, B and B' domains) are surrounded in green. Secondary structure elements belonging to the (β/{alpha})8 catalytic domain are surrounded in red ({alpha}-helixes) or blue (β-strands). The residues (D286, E328 and D393) belonging to the catalytic triad are indicated by asterisks. Amino acid mutations are surrounded in black. (A) Library B: Wild-type and variant AS sequences are aligned from residues 5 to 186 which includes the N domain and the beginning of the (β/{alpha})8 catalytic domain. (B) Library C: Wild-type and variant AS sequences are aligned over the region targeted for random mutagenesis (from residues 242 to 470) which includes the end of the B domain, a portion of the (β/{alpha})8 catalytic domain (from {alpha}-helix 3 to the beginning of {alpha}-helix 7) and the B' domain.

 

   Conclusion and future perspectives
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Conclusion and future...
Funding
 Acknowledgements
 References
 
Three large libraries of AS variants were constructed following the in vitro MutaGenTM procedure using two human error-prone DNA polymerases, pol β and pol {eta}. Unlike epPCR methods, this technique allows the generation of protein variants through a single mutagenic replication step. Moreover, sequence analysis of the libraries showed that the generated diversity was related to specific features of the employed DNA polymerases. The use of either pol β or pol {eta} allowed the creation of libraries with different mutation rates and complementary mutational spectra. The variants obtained through this approach combined all types of mutations, including deletions of codons. Hence, MutaGenTM appears as a versatile random mutagenesis tool since it takes advantage of the mutational bias of several error-prone DNA polymerases. The potential of this method will be enlarged through the use of other mutagenic polymerases like pol {iota} or pol {kappa}.


   Funding
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Conclusion and future...
 Funding
Acknowledgements
 References
 
S.E. was financially supported by MilleGen SA and l’Association Nationale de la Recherche Technologique (France).


   Footnotes
 
Edited by Alan Berry


   Acknowledgements
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Conclusion and future...
 Funding
 Acknowledgements
References
 
The authors wish to thank David Grand for the production and the purification of the DNA polymerases used in this study, Pascale Klopp and Nathalie Souyris for helpful technical assistance in the creation of libraries and the group supervised by Yann Merlet for the sequencing of variants.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Conclusion and future...
 Funding
 Acknowledgements
 References
 
Albenne C., et al. J. Biol. Chem. (2004) 279:726–734.[Abstract/Free Full Text]

Beard W.A., Wilson S.H. Structure (2003) 11:489–496.[Medline]

Bebenek K., Matsuda T., Masutani C., Hanaoka F., Kunkel T.A. J. Biol. Chem. (2001) 276:2317–2320.[Abstract/Free Full Text]

Bouayadi K., Kharrat H., Louat T., Servant L., Cazaux C., Hoffmann J.S. (2002) Patent WO0238756.

Boudsocq F., Kokoska R.J., Plosky B.S., Vaisman A., Ling H., Kunkel T.A., Yang W., Woodgate R. J. Biol. Chem. (2004) 279:32932–32949.[Abstract/Free Full Text]

Cadwell R.C., Joyce G.F. PCR Meth. Appl. (1992) 2:28–33.[Medline]

Cline J., Hogrefe H.H. Strategies (2000) 13:157–161.

Eckert K.A., Kunkel T.A. Nucleic Acids Res. (1990) 18:3739–3744.[Abstract/Free Full Text]

Flores-Ramirez G., Rivera M., Morales-Pablo A., Osuna J., Soberon X., Gaytan P. BMC Chem. Biol. (2007) 7:1.[CrossRef][Medline]

Fujii R., Kitaoka M., Hayashi K. Nucleic Acids Res. (2006) 34:e30.[Abstract/Free Full Text]

Garcia-Diaz M., Kunkel T.A. Trends Biochem. Sci. (2006) 31:206–214.[CrossRef][Web of Science][Medline]

Grishin N.V. J. Struct. Biol. (2001) 134:167–185.[CrossRef][Web of Science][Medline]

Jones D.D. Nucleic Acids Res (2005) 33:e80.[Abstract/Free Full Text]

Kunkel T.A. J. Biol. Chem. (1985) 260:5787–5796.[Abstract/Free Full Text]

Kunkel T.A. J. Biol. Chem. (2004) 279:16895–16898.[Free Full Text]

Kunkel T.A., Pavlov Y.I., Bebenek K. DNA Repair (2003) 2:135–149.[Medline]

Lantto J., Ohlin M. J. Biol. Chem. (2002) 277:45108–45114.[Abstract/Free Full Text]

Leung D.W., Chen E., Goeddel D.V. Technique (1989) 1:11–15.

Lin-Goerke J.L., Robbins D.J., Burczak J.D. BioTechniques (1997) 23:409–412.[Web of Science][Medline]

Ling H., Boudsocq F., Woodgate R., Yang W. Cell (2001) 107:91–102.[CrossRef][Web of Science][Medline]

Matsuda T., Bebenek K., Masutani C., Rogozin I.B., Hanaoka F., Kunkel T.A. J. Mol. Biol. (2001) 312:335–346.[CrossRef][Web of Science][Medline]

Matsumoto Y., Kim K. Science (1995) 269:699–702.[Abstract/Free Full Text]

Mondon P., Souyris N., Douchy L., Crozet F., Bouayadi K., Kharrat H. Biotechnol. J. (2007) 2:76–82.[CrossRef][Medline]

Moon A.F., Garcia-Diaz M., Batra V.K., Beard W.A., Bebenek K., Kunkel T.A., Wilson S.H., Pedersen L.C. DNA Repair (2007) 6:1709–1725.[CrossRef][Medline]

Murakami H., Hohsaka T., Sisido M. Nat. Biotechnol. (2002) 20:76–81.[CrossRef][Web of Science][Medline]

Osheroff W.P., Jung H.K., Beard W.A., Wilson S.H., Kunkel T.A. J. Biol. Chem. (1999) 274:3642–3650.[Abstract/Free Full Text]

Osuna J., Yanez J., Soberon X., Gaytan P. Nucleic Acids Res. (2004) 32:e136.[Abstract/Free Full Text]

Potocki de Montalk G., Remaud-Simeon M., Willemot R.M., Planchot V., Monsan P. J. Bacteriol. (1999) 181:375–381.[Abstract/Free Full Text]

Prakash S., Johnson R.E., Prakash L. Ann. Rev. Biochem. (2005) 74:317–353.[CrossRef][Web of Science][Medline]

Sarcabal P., Remaud-Simeon M., Willemot R., Potocki de Montalk G., Svensson B., Monsan P. FEBS Lett. (2000) 474:33–37.[CrossRef][Web of Science][Medline]

Shafikhani S., Siegel R.A., Ferrari E., Schellenberger V. BioTechniques (1997) 23:304–310.[Web of Science][Medline]

Silvian L.F., Toth E.A., Pham P., Goodman M.F., Ellenberger T. Nat. Struct. Biol. (2001) 8:984–989.[CrossRef][Web of Science][Medline]

Simm A.M., Baldwin A.J., Busse K., Jones D.D. FEBS Lett. (2007) 581:3904–3908.[CrossRef][Medline]

Sinha N.K., Haimes M.D. J. Biol. Chem. (1981) 256:10671–10683.[Abstract/Free Full Text]

Skov L.K., Mirza O., Sprogoe D., Dar I., Remaud-Simeon M., Albenne C., Monsan P., Gajhede M. J. Biol. Chem. (2002) 277:47741–47747.[Abstract/Free Full Text]

Streisinger G., Okada Y., Emrich J., Newton J., Tsugita A., Terzaghi E., Inouye M. Cold Spring Harb. Symp. Quant. Biol. (1966) 31:77–84.[Abstract/Free Full Text]

Taylor M.S., Ponting C.P., Copley R.R. Genome Res. (2004) 14:555–566.[Abstract/Free Full Text]

Tindall K.R., Kunkel T.A. Biochemistry (1988) 27:6008–6013.[CrossRef][Web of Science][Medline]

Trincao J., Johnson R.E., Escalante C.R., Prakash S., Prakash L., Aggarwal A.K. Mol. cell (2001) 8:417–426.[CrossRef][Web of Science][Medline]

van der Veen B.A., Potocki-Veronese G., Albenne C., Joucla G., Monsan P., Remaud-Simeon M. FEBS Lett. (2004) 560:91–97.[CrossRef][Web of Science][Medline]

Vanhercke T., Ampe C., Tirry L., Denolf P. Anal. Biochem. (2005) 339:9–14.[CrossRef][Web of Science][Medline]

Vartanian J.P., Henry M., Wain-Hobson S. Nucleic Acids Res. (1996) 24:2627–2631.[Abstract/Free Full Text]

Wong T.S., Zhurina D., Schwaneberg U. Comb. Chem. High throughput Screen. (2006) 9:271–288.[CrossRef][Web of Science][Medline]

Yuan L., Kurek I., English J., Keenan R. Microbiol. Mol. Biol. Rev. (2005) 69:373–392.[Abstract/Free Full Text]

Zaccolo M., Williams D.M., Brown D.M., Gherardi E. J. Mol. Biol. (1996) 255:589–603.[CrossRef][Web of Science][Medline]

Zhou B.L., Pata J.D., Steitz T.A. Mol. Cell (2001) 8:427–437.[CrossRef][Web of Science][Medline]

Received October 1, 2007; revised January 13, 2008; accepted January 14, 2008.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrow All Versions of this Article:
21/4/267    most recent
gzn004v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Emond, S.
Right arrow Articles by Remaud-Simeon, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Emond, S.
Right arrow Articles by Remaud-Simeon, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?