Fig 1.
Scheme of the first selection cycle.
The selection cycle consists of two types of selection: in vitro selection of more replicable RNA and in vivo functional selection of the α-domain gene of β-galactosidase. (1) DNA fragments containing the α-domain gene, the terminal recognition sites for Qβ replicase, and the T7 promoter were amplified, and mutations were introduced by error-prone PCR to obtain a DNA library. (2) The DNA library was transcribed in vitro with T7 RNA polymerase to produce the RNA library. (3) In vitro RNA replication was performed with Qβ replicase. In this step, more replicable mutant RNAs in the library were selected. (4) The replicated RNAs were reverse transcribed into cDNAs, (5) ligated to a plasmid vector and (6) transformed into an E. coli strain that expressed the ω-domain of β-galactosidase. (7) The transformed E. coli were plated on an agar plate containing the colorimetric substrate of β-galactosidase. The cells harboring the plasmids containing the functional α-domain gene produced blue-colored colonies and were collected. (8) The plasmids were then extracted from the colony mixture for the next cycle.
Fig 2.
Replication and α activities of RNA clones after the selection cycle.
A) Four RNA clones were randomly chosen after the selection cycle. These RNAs (0.1 nM) were replicated with the Qβ replicase (100 nM) for 1 h at 37°C and subjected to 8% polyacrylamide-gel electrophoresis followed by RNA staining with SYBR green II. SS and DS indicate the position of single- or double-strands, respectively. B) To obtain quantitative data, the RNA concentration of the original and m4 RNA in the replicated solutions were measured by quantitative PCR after reverse-transcription. Error bars represent standard error. C) The α-complementation activities of the α-domain genes. The plasmids encoding each RNA clone were introduced into an E. coli strain expressing the ω-domain gene, and the cells were streaked on a plate containing the colorimetric substrate for β-galactosidase. If the plasmids contain functional α-domain gene, the cell streaks become blue.
Table 1.
List of mutations.
Fig 3.
Estimation of the kinetic parameters of the original and m4 RNAs.
To estimate the kinetic parameters, KM and kcat, the rates of complementary strand synthesis were measured using the plus (A) or minus strands (B) of the original and m4 RNA clones. Each template RNA (0, 5, 10, 20, 40, and 100 nM) was incubated with the replicase (10 nM) at 37°C for 5 min, and the synthesized RNA was measured as described in the Materials and Methods section. The plots were fitted to Michaelis-Menten equations (solid lines) for estimating Vmax and KM. To estimate the kcat values, we used the active ratio of the replicase (20%) [23]. The estimated parameters are shown in Table 2 along with the fitting errors. The error bars represent standard errors (n = 3).
Table 2.
Kinetic parameters.
Fig 4.
RNA replication of reverse mutations from the m4 minus RNA.
Eight reverse mutant RNAs (10 nM), in which each of the eight mutations introduced in the m4 minus RNA were reverted to the original nucleotides, were incubated with the replicase (10 nM) for 0.5 h at 37°C. The replication amount was measured as described in the Materials and Methods. For comparison, the replication level of the original and m4 minus RNA are shown (dotted lines). The error bars represent standard errors (n = 3).
Fig 5.
Scheme of the second selection cycle.
The second selection cycle is the same as the first one (Fig 1) except for the changes in the in vivo selection method and omission of PCR amplification and mutagenesis. (1) A plasmid encoding the target gene, terminal recognition sites for Qβ replicase, and T7 promoter was transcribed in vitro with T7 RNA polymerase to produce an RNA. (2) In vitro RNA replication was performed with Qβ replicase. In this step, more replicable mutant RNAs from the library were selected. (3) The replicated RNAs were reverse transcribed into cDNAs and (4) ligated to a plasmid vector. (5) The ampicillin resistance plasmid was introduced into an E. coli strain with deleted target gene and complemented with a plasmid encoding sucB gene for negative selection. (6) After 10h incubation in the presence of ampicillin, the small colonies were transferred on a new agar plate containing 5% sucrose for selection of cells that lost the original complementary plasmid. An E. coli cell harboring the plasmid encoding the functional target gene would produce a large colony on this plate. (7) All large colonies were collected and plasmids extracted from the colony mixtures for the next cycle.
Fig 6.
Model selection experiments of serS and glnS.
A 1:1 mixture of plasmids encoding wild-type gene and non-functional mutant gene was subjected to in vivo functional selection (steps [5–7] in Fig 5). The plasmid mixture before (lane B) and after (lane A) selection was directly digested with EcoRV and PstI for glnS or digested with ClaI after PCR amplification with primers, GGCGATTAAGTTGGGTAACGCCAG and CCGGCTCGTATGTTGTGTGG for SerS. Bands specific for mutant gene were indicated with arrowheads. The “M” lane indicates λ-Hind III marker.
Fig 7.
Replication of a serS RNA clone after the combinatorial selection cycle.
The original RNA and the RNA (sR10) obtained after the combinatorial selection were incubated with replicase (20 nM) for 0.5 h at 37°C. The replication amount was measured as described under Materials and Methods. Error bars represent standard errors (n = 3).