Fig 1.
Plasmid assembly by E. coli in vivo recombination of multiple DNA fragments (F1, F2, F3, and F4) with overlapping ends.
(A) Assembly by 2 DNA fragments, (B) Assembly by 4 DNA fragments. The length of overlapping ends is indicated by X nt OL (e.g. 18 nt OL). The DNA fragments should contain a replication origin (Ori), an antibiotic resistant gene (AmpR or KanR), a promoter, and one or more target genes (or gene fragments). Each of the DNA fragments is generated by PCR from DNA templates and a pair of primers (P1/P2, P3/P4) with high fidelity Q5 DNA polymerase.
Fig 2.
DNA fragments of varying sizes (from 344–11,792 bp) prepared by PCR used for the construction of plasmids.
(A) pGFP (2.6 kb), (B) pIkB (6.3 kb), (C) pDcEG (11.9 kb), and (D) pDSADE (16.1 kb). The 1 kb band in the DNA marker lane (M, 1.5 μL NEB Fast DNA ladder) contained 8.1 ng DNA. Aliquots of 1.5 μL PCR reactions were loaded onto 0.8–1.5% agarose gels containing EtBr. After running 45–60 min at 80–100 V, the gels were scanned by a phosphorimager under Cy3 fluorescence setting. The sequences of constructed plasmids, PCR primers, and overlapping regions are included in S1–S4 Tables and S1–S6 Sequences.
Fig 3.
Colonies from 2-fragment assembly of varying lengths of overlapping nucleotides (0–25 nt OL) to construct the pGFP plasmid.
The zero background can be seen from 0 and 6 nt OL. Increasing the OL size generates a higher number of colonies.
Fig 4.
The relationship between cloning efficiency and the length of overlapping ends.
(A) Identification of positive and negative colonies by fluorescence. The circled colonies were non-fluorescent, which were used to calculate cloning accuracy of plasmid assembly by E. coli in vivo recombination. (B) The relationship between the number of colonies and overlapping nucleotides through 2-fragment assembly to construct pGFP. The data were from 2–6 independent experiments for each point, and the error bar represents the range of variation. (C) High cloning accuracy of pGFP 2-fragment assembly, regardless of the length of overlapping nucleotides. The 9 nt OL point had apparent 100% accuracy and no variation from 2 independent experiments, but it only had 7 colonies in total.
Fig 5.
Colonies from in vivo assembly of 3–5 overlapping DNA fragments with 18 nt OL and 25 nt OL to construct pGFP.
While increasing the OL size produces more colonies, the number of colonies decreases with increasing number of DNA fragments.
Fig 6.
The correlation between the number of colonies and the number of DNA fragments (2–5) with 18 nt OL and 25 nt OL to construct pGFP.
For both 18 and 25 nt OL assemblies, the colony number rapidly decreases with the number of DNA fragments. The error bar represents the range of variation.
Fig 7.
Correlation of plasmid size and the number of colonies.
(A) The relationship between the number of colonies and plasmid size through 2-fragment assembly with 18 nt OL and 25 nt OL. (B) The relationship between the number of colonies and plasmid size from 3-fragment assembly with 18 nt OL and 25 nt OL. The error bar represents the range of variation.
Fig 8.
Identification of positive colonies by colony PCR from the construction of (A) pIkB, (B) pDcEG, and (C) pDSADE. The expected correct DNA products are marked on the right sides of the gels. The smear bands below the correct DNA product bands are not consistent from different PCR reactions, but their identity and reasons for inconsistency are unknown.
Fig 9.
Confirmation of correct plasmid assembly by digestion with restriction enzymes.
(A) pIkB digestion by SalI & NdeI resulting in 3 expected bands at 842, 2508, and 2979 bp, (B) pDcEG digestion by NsiI leading to 4 expected bands at 266, 1611, 2996, and 6940 bp, and (C) pDSADE digestion by NsiI generating 5 expected bands at 266, 1751, 2996, 4048, and 6940 bp. The 266 bp band was barely visible due to its expected weak intensity relative to other bands (only 3.8% of the 6940 bp band). No effort was made to identify the single negative plasmid (lane 4 in C), but it was about 2.8 kb and contained KanR (the presence of the 266 bp signature band by NsiI digestion) as expected for growth under kanamycin conditions. It was not one of the two parental template plasmids (pDcEG & pDSA, S4 Table) used for PCR.