Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (232K) Freely available
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 ISI Web of Science
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 arrow Search for citing articles in:
ISI Web of Science (42)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Ohuchi, S.
Right arrow Articles by Yamane, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ohuchi, S.
Right arrow Articles by Yamane, T.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 4339-4346  

In vitro method for the generation of protein libraries using PCR amplification of a single DNA molecule and coupled transcription/translation
Introduction
Materials And Methods
   Oligonucleotide primers
   Plasmid construction
   Preparation of PCR templates
   PCR amplification of a single DNA molecule (single-molecule PCR)
   In vitro coupled transcription/translation
   Sequence analysis of mutations
Results
   Single-molecule PCR
   Expression of gfp genes amplified by single-molecule PCR
   Sequencing analysis of mutations on single-molecule PCR
Discussion
Acknowledgements
References

In vitro method for the generation of protein libraries using PCR amplification of a single DNA molecule and coupled transcription/translation

In vitro method for the generation of protein libraries using PCR amplification of a single DNA molecule and coupled transcription/translation

Shoji Ohuchi, Hideo Nakano* and Tsuneo Yamane

Laboratory of Molecular Biotechnology, Department of Biological Mechanisms and Functions, Graduate School of Biological and Agricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan

Received July 9, 1998; Revised and Accepted August 17, 1998

ABSTRACT

A novel in vitro method for the generation of a protein library has been developed using the polymerase chain reaction (PCR) amplification of a single DNA molecule followed by in vitro coupled transcription/translation. DNA template encoding green fluorescent protein of a jellyfish Aequorea victoria was extensively diluted to one molecule per well, and then amplified by a total of 80 cycles of PCR with nested primers. The exact number of origins in the amplified DNA fragment was then estimated by directly sequencing a part of the fragment, at which an individual template molecule was marked by PCR with a primer containing three randomized bases. Since the sequences obtained in 91 independent amplifications were diversified statistically, each amplified fragment was likely originated from a single DNA molecule. In addition, the amplified fragments served as a template for in vitro coupled transcription/translation using T7 RNA polymerase and Escherichia coli S30 extract. These results suggest that the library obtained by the PCR amplification of a single DNA molecule diluted from a variety of DNA pools is potentially useful in high-throughput generation of protein libraries.

INTRODUCTION

A new methodology generating combinatorial libraries, which has great potential for the screening of compounds with desired properties, is now in rapid expansion (1-4). The combinatorial library for the screening and selection of proteins includes a variety of biological systems such as phage display (5-8), peptides-on-plasmids library (9), cell surface display (10) and others. Although such a cell-based system has been widely used, generally its library scale is physically limited by the use of agar plates, and the screenings and selections cannot be applied to toxic proteins due to the necessity of cell growth. On the other hand, protein selection can be carried out in vitro by ribosome display (11,12) and RNA-peptide fusion display (13,14). However, such systems can be used only for selection by affinity binding.

Recently, the technology of automatic, high-speed screening and selection has progressed for combinatorial chemistry, where high-speed synthesis together with high-throughput screening (HTS) allows screening of hundreds-of-thousands to millions of compounds per day (15,16). However, organic peptide synthesis in the traditional stepwise fashion, which is usually used to generate a peptide library in combinatorial chemistry, is limited to short sequences. Here, we have developed a simple system to generate a protein library by sequentially combining the polymerase chain reaction (PCR) with in vitro transcription and translation.

In this novel system, a protein library can be generated by in vitro transcription and translation of templates amplified from a single DNA molecule diluted from a variety of DNA pool. To detect the number of template molecules in each amplification, every DNA molecule was individually marked by PCR with a primer containing three randomized bases. By directly sequencing the amplified fragments, the successful amplification of the DNA fragment from a single origin was confirmed experimentally. Mutant and wild-type green fluorescent protein (GFP) genes were then mixed once at the ratio of 1:1 or 1:9, `cloned' by the PCR, and expressed by Escherichia coli coupled transcription/translation. The fluorescent signals representing wild-type and mutant GFPs were obtained at approximately the mixed ratios. This novel method, named `PCR library', is easy to miniaturize and automate. It enables high-speed screening of proteins as well as peptides generated by combinatorial chemistry.

MATERIALS AND METHODS

Oligonucleotide primers

All oligonucleotide primers were kindly provided by Nippon Flour Mills (Tokyo, Japan). The sequences of the primers are as follows: F-1, d(GCGAGTCAGT GAGCGAGGAA G); F-2 d(CGATTCATTA ATGCAGATCT CGATCCC); R-0, d(AACAGTACAC GTCGTGCCAC CAAAGCAGAG AGCTCCACTG TCTCGCCAAN NNGATCAAGC TTCGAATTCT ACGAATGCTA); R-1, d(AACAGTACAC GTCGTGCCAC C); R-2, d(AGCAGAGAGC TCCACTGTCT CGCC); S65OL, d(GAAAGTAGTG ACAAGTGTTG GC); S65TOL, d(GCCAACACTT GTCACTACTT TCACCTATGG TGTTCAATGC TTTTC); F-T7TER-1, d(GGCGAGACAG TGGAGCTCTC TGCCCGGCTG CTAACAAAGC C); R1-32, d(CTGCGCAACT GTTGGGAAGG G).

Plasmid construction

The E.coli strain XL1-Blue (Stratagene) was used as the host for all DNA manipulations. The gfp gene of pGFP (Clontech) was isolated by the digestion of BamHI and EcoRI, and ligated to BamHI/EcoRI-digested pRSETB (Invitrogen), resulting plasmid pRSET-GFP1.

The 5[prime] terminal part of gfp gene was amplified using a set of primers of F-1 and S65OL to replace Ser to Thr at position 65, which affects the excitation wavelength. The 3[prime] part was also amplified using S65TOL and R-0, and the two fragments were joined by overlap extension PCR and digested by HincII and NcoI. Then the 0.32-kb fragment containing the mutation site was ligated to HincII/NcoI-digested pRSET-GFP1, resulting plasmid pRSET-GFP2. The sequence of the inserted fragment was confirmed by the dye-terminator sequencing method.

Preparation of PCR templates

To prepare the target DNA for PCR amplification of a single DNA molecule, the ScaI digest of pRSET-GFP1 was first amplified by 30 cycles of conventional PCR using Taq DNA polymerase (Takara Shuzo, Kyoto, Japan) with 1 µM of primers F-1 and R-0 containing three randomized bases under the reaction conditions recommended by the supplier.

The PCR product was extracted with phenol-chloroform, precipitated with ethanol, and purified by ion-exchange HPLC on a TSKgel DEAE-NPR column (Tosoh, Tokyo, Japan) in a gradient of NaCl in 20 mM Tris-HCl (pH 9.0) as follows: flow rate = 0.5 ml/min; 0 M NaCl at 0 min; 0.5 M NaCl at 10 min; 0.7 M NaCl at 30 min; 1 M NaCl at 50 min.

In order to remove heteroduplex sequences that were formed by extension of the primer containing a random sequence (R-0) at the final cycle of PCR, five cycles of PCR were additionally carried out using the chromatographically purified template and primer R-1 instead of R-0. The product was then purified again by ion-exchange HPLC.

The target DNAs corresponding to wild-type and S65T GFP for in vitro expression were prepared from a ScaI digest of pRSET-GFP1 and pRSET-GFP2, respectively, by conventional PCR using Pfu DNA polymerase (Stratagene) with primers (F-1 and R-0) in the following sequence: preheat at 94°C for 3 min; 25 cycles consisting of 96°C for 15 s, 55°C for 30 s, and 72°C for 3 min; additional extension at 72°C for 10 min. The PCR products were purified by ion-exchange HPLC as above. The purified products were amplified by five additional cycles of PCR as described above.

PCR amplification of a single DNA molecule (single-molecule PCR)

After the target DNA was diluted with 1 ng/µl tRNA (Sigma) as a carrier typically to one molecule per well (DNA concentrations were determined by measuring absorbance at 260 nm, and1.18 × 10-6 ng of template DNA molecule was estimated as one molecule), each molecule was first amplified in 10 µl of total volume by the 0.4 U Taq DNA polymerase in the recommended reaction condition with 0.2 mM of each dNTP, and 0.1 µM of each primer (F-1 and R-1) in the following sequence: preheat at 94°C for 3 min; 50 cycles consisting of 96°C for 15 s, 53°C for 30 s and 72°C for 1 min; additional extension at 72°C for 10 min. One µl of the first-step PCR product was transferred to 10 µl of fresh PCR mixture containing 0.4 U Taq DNA polymerase, 0.2 mM of each dNTP and 0.5 µM of each primer (F-2 and R-2), and then processed as follows: preheat at 94°C for 3 min; 30 cycles consisting of 96°C for 15 s, 58°C for 30 s and 72°C for 1 min; additional extension at 72°C for 10 min.

Before in vitro coupled transcription/translation, a T7-termination sequence was added to the PCR products by the overlap extension PCR method. The T7-termination fragment was amplified by Taq DNA polymerase with primers F-T7TER-1 and R1-32 from ScaI digest of pRSET-GFP1 as a template. After HPLC-purification, the T7-termination fragment was attached to the product of single-molecule PCR by the overlap extension PCR. The product (0.1 µl) of single-molecule PCR was used as a template for the overlap extension PCR, with primers F-2 and pRSET-1-32. The reaction (30 cycles) was carried out in 10 µl of total volume by the Taq DNA polymerase.

In vitro coupled transcription/translation

The in vitro coupled transcription/translation reaction was carried out essentially by Schultz's method (17) with some modifications (T. Endo, Nagoya University, Japan). T7 RNA polymerase and E.coli S30 extract for the reaction were prepared using standard procedures. NTPs were purchased from Pharmacia, creatine phosphate from Calbiochem, creatine kinase and E.coli tRNA from Boehringer Mannheim, folinic acid and rifampicin from Sigma. Polyethylene glycol (Mr 6000) and other chemicals were from Wako Pure Chemical Industries (Osaka, Japan).

The reaction was carried out in 20 µl total volume, and the reaction mixture contained the following: 10 mM magnesium acetate, 56.4 mM Tris-acetate (pH 7.4), 1.76 mM DTT, 1.22 mM ATP, 0.85 mM GTP, 0.85 mM CTP, 0.85 mM UTP, 40 mM creatine phosphate, 0.15 mg/ml creatine kinase, 0.32 mM of each amino acid, 4% polyethylene glycol (Mr 6000), 34.6 mg/ml folinic acid, 0.17 mg/ml E.coli tRNA, 150 mM potassium acetate, 35.9 mM ammonium acetate, 3 µl E.coli S30 extract, 10 mg/ml T7 RNA polymerase, 10 mg/ml rifampicin. Reactions were initiated by adding 2 µl of the overlap extension PCR products or 0.4 µg of plasmid DNA. Reactions were incubated at 37°C for 2 h and stopped by placing on ice.

Reaction mixtures (10 µl) were diluted with an appropriate volume of 50 mM sodium phosphate buffer (pH 7.0), and excitation spectra of the diluted mixtures were measured by a fluorescence spectrophotometer F-4500 (Hitachi). Measurements were carried out in a 1 ml cell, with a fixed emission of 510 nm. The emission monochro-metor was set at a band pass of 5 nm and scan speed was 60 nm/min.

Sequence analysis of mutations

A part of gfp gene used for in vitro coupled transcription/translation reaction was cloned into the E.coli strain XL1-Blue and sequenced. One of the DNA fragments amplified from a single origin which showed the excitation spectrum of only wild-type GFP was digested by EcoRI and EcoT14I. A 0.56 kb fragment containing a part of gfp was purified from agarose gel. The fragment was then ligated to EcoRI/EcoT14I-digested pTV119N (Takara Shuzo), transformed, and sequenced by the dye-terminator sequencing method.

RESULTS

Single-molecule PCR

Briefly, we attempted in vitro amplification of a single DNA molecule by PCR, tentatively termed `single-molecule PCR'. Under such an extreme condition, the number of template DNA molecules was so small that the formation of byproducts such as primer dimers, even if only a few were generated, might strongly inhibit the amplification of the desired products. To reduce the effect of the byproducts, therefore, we adopted the nested PCR method (Fig. 1) for amplification (18,19). The nested PCR consisted of a first-step amplification with outside primers and a second-step amplification with inside (nested) primers (Fig. 1). In the second-step reaction, nested primers should anneal with and amplify only the desired product.


Figure 1. Schematic diagram of the strategy for single-molecule PCR. Step 1: a pool of template DNA is generated by PCR, whereby three randomized bases are introduced downstream of gfp gene in order to distinguish each template. The fragment contains T7 promoter for in vitro expression. Step 2: the generated DNA pool is purified and diluted to the specified concentration. Step 3: first step of amplification is carried out by 50 cycles of PCR with a set of outside primers. Step 4: second step of amplification is carried out by 30 cycles of PCR with a set of inside (nested) primers.

Does an amplified fragment have a single origin? Could a single DNA molecule really be amplified? In order to confirm this point, we distinguished each template by virtue of introduction of three randomized bases into the template. The template DNA prepared by PCR carries gfp gene, T7-promoter and the three randomized bases (Fig. 1). There were 64 possible sequences at the three bases (4 × 4 × 4). If an amplified fragment has a single origin, only a unique sequence should be seen in the three randomized bases of the fragment. By directly sequencing the three bases of the fragments obtained in each amplification, therefore, the number of template molecules can be estimated.

The template DNA thus prepared was diluted to one molecule per well, followed by the nested PCR as described. Optimized concentrations of the primers used in the nested PCR were 0.1 and 0.5 µM in the first and second step of the amplification, respectively. To avoid annealing of the first-step primers at the second-step PCR, the second-step primers were designed to have higher melting temperatures than those of the first-step primers. Before sequencing, it was confirmed by agarose gel electrophoresis that the correct size of a DNA fragment was amplified.

The experiments were carried out seven times, and the results are summarized in Table 1. None of the 16 negative (no template) controls gave a false-positive signal. Successful amplification was observed at frequencies of 68.7, 48.9, 27.9 and 23.8% at 4, 2, 1 and 0.5 molecules of template per well, respectively. On the other hand, based on the Poisson distribution, the theoretically calculated frequencies of amplifiable (positive) wells at 4, 2, 1 and 0.5 DNA molecules per well are 98.2, 86.5, 63.2 and 39.4%, respectively.

Following the agarose gel analysis, the amplified fragment from the PCR mixture was sequenced directly. Figure 2A and B shows typical signals obtained by sequencing the amplified fragment in the experiments of single-molecule PCR, and C shows that obtained from ~1 × 104 molecules of template DNA as control. At the three randomized bases, A shows a clear signal that exclusively indicates a unique sequence, while C shows an unclear signal that reflects its numerous number of origins. Although B shows a clear signal, it is likely that at least two sequences overlapped, indicating that the amplified fragment had more than one origin.

Table 2 shows all sequences obtained at the three randomized bases of the amplified fragments from a single DNA molecule. Although 92 independent sequences were obtained, 91 of them were shown in Table 2A, because one gave only two bases (CC). From a theoretical calculation, 48.6 kinds of sequences should be filled by 91 independent sequences. Since a comparable number of the variety of sequences, 42, was obtained experimentally, the possibility of contamination such as carry-over can be excluded.

Expression of gfp genes amplified by single-molecule PCR

In order to demonstrate the usefulness of the PCR library for in vitro screening, direct expression of the gfp gene amplified from a single molecule was examined by in vitro coupled transcription/translation by means of T7 RNA polymerase and E.coli S30 extract. In addition to the wild-type GFP (WT), an S65T mutant GFP (S65T) was used, which has greater brightness than WT and shifted the excitation maximum to 488 nm rather than 396 nm of WT (20). These two are easily distinguishable from each other. To avoid the introduction of mutations, a high-fidelity DNA polymerase, Pfu DNA polymerase, was used rather than Taq polymerase for the template preparation.

The gfp genes of WT and S65T were amplified using Taq DNA polymerase as described above, except for the addition of TaqStart Antibody (Clontech) to carry out hot-start PCR. The template DNAs were diluted statistically to one or two molecules per well, then the nested PCR was carried out in all 64 wells: 16 wells with two DNA molecules of WT; 16 with one molecule of WT; 16 with two molecules of S65T; 16 with one molecule of S65T. In each template condition, successful amplification was observed at 4, 3, 3 and 3 wells, respectively. The T7-termination sequence was then added to the amplified fragments by the overlap extension PCR, because the transcription efficiency of the PCR product could be increased by the addition of the sequence (to be published elsewhere).

Table 1. Summary of single-molecule PCR
Experiment Input DNA
(molecules/well)		 Number of wells Number of positive wells Single moleculea Multiple moleculesa N.D.b
1 4 16 3 (18.8%) 2 1 0
2 16 2 (12.5%) 2 0 0
1 16 3 (18.8%) 1 2 0
0 s8 0 (0%) 0 0 0
2 4 16 7 (43.8%) 4 1 2
2 16 3 (18.8%) 3 0 0
1 16 2 (12.5%) 2 0 0
0 8 0 (0%) 0 0 0
3 4 15 10 (66.7%) 1 9 0
2 16 14 (87.5%) 3 8 3
1 16 7 (43.8%) 5 2 0
4 4 16 13 (81.3%) 6 7 0
2 16 10 (62.5%) 4 6 0
1 16 2 (12.5%) 2 0 0
0.5 8 1 (12.5%) 0 1 0
5 4 24 18 (75.0%) 3 13 2
2 23 13 (56.5%) 8 4 1
1 24 8 (33.3%) 7 0 1
0.5 24 5 (20.8%) 4 1 0
6 4 23 21 (91.3%) 7 11 3
2 24 14 (58.3%) 3 8 3
1 24 9 (37.5%) 4 5 0
0.5 24 9 (37.5%) 7 2 0
7 4 24 20 (83.3%) 5 13 2
2 24 10 (41.7%) 4 5 1
1 24 7 (29.2%) 3 4 0
0.5 24 4 (16.7%) 2 1 1
Total 4 134 92 (68.7%)  
2 135 66 (48.9%)
1 136 38 (27.9%)
0.5 80 19 (23.8%)
0 16 0 (0%)
aSingle molecule and multiple molecules show number of positive wells in which amplified fragment has a single origin and has multiple origins, respectively.
bNot determined.

All of the 13 gfp genes thus modified were directly expressed by in vitro coupled transcription/translation system. The excitation spectrum of the synthesized GFP was measured with an emission of 510 nm. The samples from WT gene and those from S65T showed an excitation maximum around 396 nm (a WT phenotype), and around 488 nm (an S65T phenotype), respectively (data not shown).

In order to mimic a screening procedure, we next amplified WT and S65T genes diluted from a mixed template pool, and expressed the amplified gene. One DNA pool contained equal amounts of WT and S65T, and the other a 9-fold excess of WT. Each of the template pools was diluted statistically to four, two or one molecule per well, then the nested PCR was carried out in all 96 wells; 16 wells in each template condition. The amplifications and the sequences of the three randomized bases were analyzed as described above and are summarized in Table 33. From the results in Table 33, we classified the 53 samples into four groups; group I, 20 samples with multiple origins obtained from the pool of WT:S65T = 1:1; group II, six samples with a single origin from WT:S65T = 1:1; group III, 13 samples with multiple origins from WT:S65T = 9:1; group IV, 14 samples with a single origin from WT:S65T = 9:1.


Figure 2. (A and B) Typical sequencing signals obtained by single-molecule PCR. (C) Obtained from ~1 × 104 molecules of template DNA as control. The positions of the three randomized bases are indicated by arrows.

The above 53 gfp genes were directly expressed in vitro after the addition of the T7-termination sequence. In group I, eight samples were S65T phenotype, 10 were WT phenotype and two showed ambiguous spectra (Fig. 3A, indicated by arrows). The ambiguous spectra with two weaker peaks of WT (at 396 nm) and S65T (at 488 nm) were also seen in group III (two were ambiguous and the others were WT, Fig. 3C) but not in group II (Fig. 3B) nor in group IV (Fig. 3D), indicating that both WT and S65T genes were the origins of the samples showing the signals. In group II, four samples were S65T phenotype and two were WT phenotype (Fig. 3B), and all samples were WT in group IV (Fig. 3D).

Table 2. (A) All sequences at the three randomized bases of the amplified fragments from a single DNA molecule
Second position First position Third position
T G C A
A 3 1 2 0 A
4 0 3 4 C
1 0 2 0 G
1 1 2 0 T
C 2 1 1 6 A
3 0 3 2 C
2 0 2 4 G
5 1 2 3 T
G 2 1 0 0 A
1 1 0 0 C
2 0 0 0 G
2 1 0 0 T
T 0 1 2 6 A
1 0 4 1 C
0 2 1 1 G
0 0 0 1 T
Table 2. (B) Frequency of nucleotides in the three randomized bases
  1Na 2Na 3Na Total
A 30.77 26.37 30.77 29.30
C 26.37 40.66 29.67 32.23
G 10.99 10.99 18.68 13.55
T 31.87 21.98 20.88 24.91
a1N, 2N and 3N show percentage of nucleotides in the first position, in the second position, and in the third position, respectively, of the three randomized bases.

Sequencing analysis of mutations on single-molecule PCR

Throughout the study, we used Taq DNA polymerase with no proofreading activity for a total of 80 cycles of single-molecule PCR. Furthermore, an additional 30 cycles of PCR should be carried out to add the T7-termination sequence to the amplified genes for effective expression. After so many cycles of PCR, large populations of the amplified fragments might contain serious mutations. In order to clarify this point, one of the gfp genes of group IV was cloned and sequenced. The cloned gene was 548 bp from 147 bp downstream of the initiation codon to the stop codon of gfp, including the fluorophore encoding region. Thirteen independent clones were recovered, and both strands of the inserts were sequenced. Seven clones had one or two nucleotide substitutions and the others had no mutation (Table 44). Two silent mutations and seven amino acid substitutions were detected out of a total of the nine mutations.

DISCUSSION

Amplification of DNA fragments even from a single cell have been reported by several groups. Successful amplification was also reported from a single molecule of genomic DNA (19,21,22). Amplification has not yet been reported from a single molecule of plasmid or PCR product, however. This study is therefore the first report demonstrating the possibility of the amplification of a single molecule of PCR product.

Because one double-stranded DNA molecule of 1 kb is only1 × 10-6 pg (note that one copy of a haploid human genome is 3 pg), it is difficult to control the number of such DNAs in a manually sequential dilution. In fact, Table 1 shows that the frequency of successful amplification was not reproducible(i.e., the ratio of positive wells at one DNA molecule per well ranged from 12.5 to 43.8%). It is also noteworthy that the frequency at every trial was less than that of the theoretical calculation from the Poisson distribution. It is not likely that the adsorption of DNA molecule on pipettes or tubes accounts for this difference, because tRNA was used as a carrier to avoid it. These facts imply that a simple comparison between the amplifiable number obtained from experiments and the theoretical calculation is not enough to demonstrate the amplification of a single DNA molecule.


Figure 3. Excitation spectra of GFP synthesized in vitro with an emission of 510 nm. (A) Group I, the sample from WT:S65T = 1:1 which has multiple origins (20 samples). (B) Group II, from WT:S65T = 1:1 which has a single origin (six samples). (C) Group III, from WT:S65T = 9:1 which has multiple origins (13 samples). (D) Group IV, from WT:S65T = 9:1 which has a single origin (14 samples). FI 510 nm is relative fluorescence intensity with an emission of 510 nm.

To obtain more direct evidence, we constructed a template DNA carrying three randomized bases to distinguish each template (Fig. 1). Direct sequencing of the amplified fragments revealed the exact number of their origins (Fig. 2). Although there is a possibility that all of the templates in the same well have the same sequence, such a rare occurrence [1/64 of two templates and (1/64)2 of three templates, and so on] can be ignored. Table 2 shows all sequences obtained at the three randomized bases. Forty-two kinds of sequences obtained from 91 samples proved to have a single origin. The number of the variety of the sequences obtained, 42, is smaller (although comparable) to that of the theoretical calculation, 48.6. This is probably caused by the inequality in chemical synthesis among nucleoside phosphoramide in the synthesis of oligonucleotide R-0, since the ratio of guanine in the three randomized bases was significantly smaller than the others (Table 2B).

Since it is impossible to know the actual number of DNA molecules in each tube, it is also impossible to distinguish between the amplification of unique DNA molecule in a compartment and that of `selected' one molecule from two or more. As summarized in Table 1, the number of DNA fragment amplified from one origin was sometimes larger than expected especially at the dilution of 4 molecule per well. In such conditions of dilution, most of the amplified fragment was probably selected from several molecules, possibly by chance, because no bias was observed in the obtained sequences as summarized in Table 2. At more dilute situation,i.e. 1 or 0.5 molecule per well, however, the chance of `selection' should be greatly reduced and hence most of the amplified fragments were likely the result of the amplification of a single molecule.

In order to demonstrate the utility of the single-molecule PCR to in vitro screening, genes encoding a wild-type (WT) and a mutant GFP (S65T) were amplified by the single-molecule PCR and expressed by E.coli S30 coupled transcription/translation using T7 RNA polymerase. Each sample showed each excitation wavelength as expected. In addition, DNA pools where WT gfp was mixed with an equal or 9-fold excess amount of S65T gfp were also targeted as above in order to mimic a screening procedure. The ratio of the excitation wavelength of GFP expressed reflected that of the gene mixed (Fig. 3). Furthermore, we could confirm that S65T expressed showed not only different excitation wavelength but also greater brightness than WT did. These results indicate the ability of PCR library for quantitative screening.

Table 3. Summary of single-molecule PCR for expression
Input DNA
(molecules/well)		 Number
of wells		 Number of
positive wells		 Single
moleculea		 Multiple
moleculesa		 N.D. b
WT:S65T = 1:1
        4 16 13 (81.3%) 4 5 4
        2 16 9 (56.3%) 1 4 4
        1 16 4 (25.0%) 1 2 1
WT:S65T = 9:1
        4 16 13 (81.3%) 4 9 0
        2 16 8 (50.0%) 5 2 1
        1 16 6 (37.5%) 5 1 0
aSingle molecule and multiple molecules show number of positive wells in which amplified fragment has a single origin and multiple origins, respectively.
bNot determined; these samples are classified into `multiple molecules' in Figure 3.

Table 4. Sequencing analysis of mutations on single-molecule PCR
Clone number Nucleotide change Amino acid change
#1 A781G M->V
#2 T414C silent
#3 A674G D->G
A669G silent
#4 T419C V->A
#5 - -
#6 A520G N->D
#7 - -
#8 - -
#9 T647C I->T
#10 - -
#11 - -
#12 A734G H->R
T377C I->T
#13 - -

In this study, a T7-termination sequence had to be added to the templates before their expression, because the productivity was too small to detect the product by fluorescence without the sequence (data not shown). This step can, however, be avoided by using a PCR template carrying the termination sequence.

Taq DNA polymerase with no proofreading activity was used for single-molecule PCR. The error rate of Taq DNA polymerase is reported to be ~0.8 × 10-5/base/cycle (23), which theoretically results in a large population of amplified DNAs containing base replacements. In order to clarify this point, one of the gfp genes used for expression was cloned and sequenced. In 13 individual clones, 548 bp fragments sequenced were found to contain a total of nine base replacements (Table 4). Accordingly, each 798 bp open reading frame of gfp is calculated to contain one base replacement on average. Assuming that the concentration of the final PCR product used for expression is 0.2 µg/µl, the error rate calculated from the result is 2.5 × 10-5/base/cycle. This value is higher than the published error rate, probably because of unfavorable amplification conditions.

Most of the mutations during single-molecule PCR were negligible in phenotypic expression of at least such a compact protein like GFP. Indeed, all of WT and S65T gfp amplified and expressed in vitro were their original phenotypes. This is because most of these random mutations are thought to be neutral (24,25). To identify an effective mutation from a selected DNA pool, however, in vivo cloning and sequencing of ~10 DNA molecules will be eventually inevitable. Therefore to reduce this task or to apply our system for larger or more intolerant proteins, the optimization of the amplification conditions is necessary to decrease the frequency of base replacements. Alternatively, other high-fidelity DNA polymerases such as Pfu DNA polymerase (23) and KOD DNA polymerase (26) have to be used. Smith and Modrich developed a method for the removal of mutant sequences from PCR products using bacterial mismatch repair proteins (27). This approach is also useful to eliminate mutations. Furthermore, it is noteworthy that miniaturization of PCR not only increases the number of samples but also reduces the cycle number of PCR, or the chance of mutation, as mentioned by Kalinina et al. (22).

The amount of protein produced by an in vitro protein synthesis system is another key factor for successful application to PCR libraries. The E.coli system that we used here has been improved greatly (28). In addition, the eukaryotic protein synthesis system using wheat germ extract has also been improved in its productivity by us (29,30). It is likely that these improvements will ensure the utility of the PCR library as the template of a protein library.

Burks et al. developed an in vitro method for rapid scanning of saturation mutagenesis (31). However, it is limited to site-specific mutations. In contrast, the method that we described here can be employed for any fashions of in vitro mutagenesis, including error-prone PCR (32,33) and DNA shuffling (34,35). Thereby the PCR library followed by in vitro protein synthesis can be used as a library for directed evolution in order to obtain an improved protein for a particular purpose (36).

One of the merits of our system is that its library scale can be greatly enlarged by miniaturization of PCR. At present, only 384-well plates are commercially available; however, its sample number can be multiplied if chip technology is applied. In addition, in vitro protein synthesis following the PCR library enables the production of toxic proteins for cells and the incorporation of non-natural amino acids (17), which greatly increases the variety of amino acid sequences in proteins.

ACKNOWLEDGEMENTS

We thank Prof. Toshiya Endo and Dr Satoshi Sekiguchi for kindly providing all the techniques needed for the E.coli coupled transcription/translation system and for the synthesis of DNA primers, respectively. This study was financially supported in part by a Grant-in-Aid for Scientific Research (No. 09650872) from the Ministry of Education, Science, Sports and Culture of Japan, and the `Research for the Future' Program of The Japan Society for the Promotion of Science.

REFERENCES

1. Schultz,J.S. and Schultz,J.S. (1996) Biotechnol. Prog., 12, 729-743. MEDLINE Abstract

2. Gallop,M.A., Barrett,R.W., Dower,W.J., Fodor,S.P.A. and Gorden,E.M. (1994) J. Med. Chem., 37, 1233-1251. MEDLINE Abstract

3. Gorden,E.M., Barrett,R.W., Dower,W.J., Fodor,S.P.A. and Gallop,M.A. (1994) J. Med. Chem., 37, 1385-1401.

4. Brown,D. (1996) Mol. Diversity, 2, 217-222.

5. Smith,G.P. (1985) Science, 228, 1315-1317. MEDLINE Abstract

6. Cortese,R., Monaci,P., Luzzago,A., Santini,C., Bartoli,F., Cortese,I., Fortugno,P., Galfr,G., Nicosia,A. and Felici,F. (1996) Curr. Opin. Biotechnol., 7, 616-621. MEDLINE Abstract

7. Dunn,I.S. (1996) Curr. Opin. Biotechnol., 7, 547-553. MEDLINE Abstract

8. O'Neil,K.T. and Hoess,R.H. (1995) Curr. Opin. Struct. Biol., 5, 443-449. MEDLINE Abstract

9. Cull,M.G., Miller,J.F. and Schatz,P.J. (1992) Proc. Natl Acad. Sci. USA, 89, 1865-1869. MEDLINE Abstract

10. Georgiou,G., Stathopoulos,C., Daugherty,P.S., Nayak,A.R., Iverson,B.L. and Curtiss,R., III (1997) Nat. Biotechnol., 15, 29-34. MEDLINE Abstract

11. Kawasaki,G. (1991) PCT Int. Appl., WO 91/05058.

12. Mattheakis,L.C., Bhatt,R.R. and Dower,W.J. (1994) Proc. Natl Acad. Sci. USA, 91, 9022-9026. MEDLINE Abstract

13. Nemoto,N., Miyamoto-Sato,E., Husimi,Y. and Yanagawa,H. (1997)FEBS Lett., 414, 405-408. MEDLINE Abstract

14. Roberts,R.W. and Szostak,J.W. (1997) Proc. Natl Acad. Sci. USA, 94, 12297-12302. MEDLINE Abstract

15. Persidis,A. (1998) Nat. Biotechnol., 16, 488-489. MEDLINE Abstract

16. Broach,J.R. and Thorner,J. (1996) Nature, 384, 14-16. MEDLINE Abstract

17. Ellman,J., Mendel,D., Anthony-Cahill,S.J., Noren,C.J. and Schultz,P.G. (1991) Methods Enzymol., 202, 301-337. MEDLINE Abstract

18. Kaneko,S., Feinstone,S.M. and Miller,R.H. (1989) J. Clin. Microbiol., 27, 1930-1933. MEDLINE Abstract

19. Holding,C. and Monk,M. (1989) Lancet, 2, 532-535. MEDLINE Abstract

20. Heim,R., Cubitt,A.B. and Tsien,R.Y. (1995) Nature, 373, 663-664. MEDLINE Abstract

21. Coutelle,C., Williams,C., Handside,A., Hardy,K., Winston,R. and Williamson,R. (1989) BMJ, 299, 22-25. MEDLINE Abstract

22. Kalinina,O., Lebedeva,I., Brown,J. and Silver,J. (1997) Nucleic Acids Res., 25, 1999-2004. MEDLINE Abstract

23. Cline,J., Braman,J.C. and Hogrefe,H.H. (1996) Nucleic Acids Res., 24, 3546-3551. MEDLINE Abstract

24. Rennell,D., Bouvier,S.E., Hardy,L.W. and Poteete,A.R. (1991) J. Mol. Biol., 222, 67-87. MEDLINE Abstract

25. Wen,J., Chen,X. and Bowie,J.U. (1996) Nat. Struct. Biol., 3, 141-148. MEDLINE Abstract

26. Takagi,M., Nishioka,M., Kakihara,H., Kitabayashi,M., Inoue,H., Kawakami,B., Oka,M. and Imanaka,T. (1997) App. Environ. Microbiol., 63, 4504-4510.

27. Smith,J. and Modlich,P. (1997) Proc. Natl Acad. Sci. USA, 94, 6847-6850.

28. Kim,D.-M., Kigawa,T., Choi,C.-Y. and Yokoyama,S. (1996) Eur. J. Biochem., 239, 881-886. MEDLINE Abstract

29. Kawarasaki,Y., Kawai,T., Nakano,H. and Yamane,T. (1995) Anal. Biochem., 226, 320-324. MEDLINE Abstract

30. Nakano,H., Tanaka,T., Kawarasaki,Y. and Yamane,T. (1996) J. Biotechnol., 46, 275-282.

31. Burks,E.A., Chen,G., Georgiou,G. and Iverson,B.L. (1997) Proc. Natl Acad. Sci. USA, 94, 412-417. MEDLINE Abstract

32. Leung,D.W., Chen,E. and Goeddel,D.V. (1989) Technique, 1, 11-15.

33. Caldwell,R.C. and Joyce,G.F. (1992) PCR Methods Appl., 2, 28-33.

34. Stemmer,W.P.C. (1994) Nature, 370, 389-391. MEDLINE Abstract

35. Stemmer,W.P.C. (1994) Proc. Natl Acad. Sci. USA, 91, 10747-10751. MEDLINE Abstract

36. Kuchner,O. and Arnold,F.H. (1997) Trends Biotechnol., 15, 523-530. MEDLINE Abstract

*To whom correspondence should be addressed. Tel: +81 52 789 4143; Fax: +81 52 789 4145; Email: hnakano@agr.nagoya-u.ac.jp

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 19 Sep 1998
Copyright©Oxford University Press, 1998.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Nucleic Acids ResHome page
T. B. Yehezkel, G. Linshiz, H. Buaron, S. Kaplan, U. Shabi, and E. Shapiro
De novo DNA synthesis using single molecule PCR
Nucleic Acids Res., July 30, 2008; (2008) gkn457v1.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
T. Kojima, Y. Takei, M. Ohtsuka, Y. Kawarasaki, T. Yamane, and H. Nakano
PCR amplification from single DNA molecules on magnetic beads in emulsion: application for high-throughput screening of transcription factor targets
Nucleic Acids Res., October 6, 2005; 33(17): e150 - e150.
[Abstract] [Full Text] [PDF]

Home page
 J. Bacteriol.Home page
J. Yang, K. Kobayashi, Y. Iwasaki, H. Nakano, and T. Yamane
In Vitro Analysis of Roles of a Disulfide Bridge and a Calcium Binding Site in Activation of Pseudomonas sp. Strain KWI-56 Lipase
J. Bacteriol., January 15, 2000; 182(2): 295 - 302.
[Abstract] [Full Text]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (232K) Freely available
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 ISI Web of Science
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 arrow Search for citing articles in:
ISI Web of Science (42)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Ohuchi, S.
Right arrow Articles by Yamane, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ohuchi, S.
Right arrow Articles by Yamane, T.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?