Single-stranded overlaps annealing generated via nicking enzymes digestion was performed as described in Figure 1A. First, target DNA fragments were amplified with a pair of specific primers. The primers consist of an overlapping sequence, a NEase site and a homologous sequence of the target DNA fragments. Because the long non-homologous sequence was placed in the primer, two cycles of PCR were performed as described in Materials and Methods. Following PCR amplification and DNA purification, all fragments were digested by nicking endonucleases to produce a nick at one single strand of the double-stranded DNA fragments, followed by incubating the digested fragments in a thermo-cycler for formation of single-stranded DNA overlaps via denaturing the double strands DNA at 90°C for 5 min. The concentrations of all the fragments were tested by Nano Drop Spectrophotometer ND-2000 after the incubation. Equal molar DNA fragments containing single-stranded overlaps were mixed together and incubated at 37°C for one hour to anneal all single-stranded overlaps together. During the in vitro annealing process, different buffers such as Fast-pfu polymerase buffer, T4 DNA polymerase and T4 DNA ligase buffers were screened. The T4 DNA ligase buffer presented the highest efficiency for all ss DNA annealing during the incubation (data not shown).

After the assembly of all DNA fragments in vitro, the in vivo circularization was performed (Fig. 1B). Two separate fragments were designed as the essential vector skeleton containing an attL site with a R6Kγorigin (R6Kγori) of replication and an attR site with a kanamycin gene (Fig. 1B). Each of these two fragments was amplified with one end containing a NEase site and an overlapping DNA (Fig. 1B), which was used to anneal with the other DNA fragments. The attL and attR sites of the essential vector skeleton can achieve site-specific recombination with the expression of integrase (Int) and excisionase (Xis). Therefore, the linear system (Fig. 1A) of the assembled fragments was circularized through attL/attR site-specific recombination. The circulation process was used in E. coli EC100D pir-116 competent cells harboring plasmid pAH83CI for Int/Xis expression (Fig. 1B and Table 1). All the primers used for polymerase chain reaction (PCR) are listed in Table 2.

In order to increase denaturation efficiency, competing primers which are identical to the ss DNA overhang were added to a denaturation mixture to function as the competitor for the removal of the complementary strands. However, there was no positive effect observed. In order to remove the competing primers and complementary strands, gel electrophoresis was also employed without positive effect observed either.

With the USER™ cloning method, the dU is excised from the PCR products only at 5′-end of the fragments as dU is designed in each PCR primer [14]. After the USER enzyme digestion, PCR products are flanked by 3′-end ss DNA extensions. Nicking enzymes digestion was conducted either at 3′-end or 5′-end to produce either 5′-end ss DNA or 3′-end ss DNA overlaps (Fig. 2). The Nt.BbvCI cassette (see Materials and Methods) was designed for 5′-end digestion to form 3′-end ss DNA overlaps, while the Nt.BspQI cassette was designed for 3′-end digestion to generate 5′-end ss DNA overlaps (Fig. 2B). The annealing results showed that 3′-end ss DNA overlaps annealed better with other fragments than the 5′-end ss DNA overlaps did, which was attributed to dephosphorylation of the first base at the 5′-end of the primers.

In order to determine the optimal overhang lengths, 5 bp, 10 bp, 15 bp, 20 bp and 25 bp ss overhangs were studied for both 3′-end and 5′-end ss DNA overlaps annealing, respectively. Results of both electrophoresis and chemical transformation showed that the overhangs of 10 bp/15 bp were able to produce the highest cloning efficiency (Fig. 2B and Table 3).

T4 DNA ligase was added in order to investigate how it affected the annealing efficiency. When added to an annealing mixture, ligation can be achieved at the annealed single-stranded DNA ends. Since our fragments were prepared by PCR amplification, the 5′-end of the PCR product lacked the phosphate and only 5′ recessed substrate has the phosphate group for ligation. Therefore, when comparing the effect of T4 DNA ligase on 5 bp, 10 bp, 15 bp, 20 bp or 25 bp ss DNA overhangs for 3′-end annealing (5′-end digestion), both results of electrophoresis and transformation revealed that the 5 bp and 10 bp were the best in the presence of T4 DNA ligase. However, in the absence of T4 DNA ligase, 10 bp and 15 bp were found to be optimal both for 3′-end and 5′-end annealing (Fig. 2B). This demonstrated that T4 DNA ligase improved the ligation for short length DNA fragments as T4 DNA ligase favors to link short sticky ends digested by Type II restriction enzymes [8]. For longer single-stranded DNA extensions, the homologous annealing was preferred compared with short single-stranded DNA extension, and both 3′-end as well as the 5′-end ss DNA overlaps annealing results demonstrated the length of a homologous tail of 10 bp and 15 bp had the highest efficiency (Fig. 2B).

Most of the LIC methods rely on the single-stranded overlaps annealing [9]. Since the lengths distribution of ss DNA overlapping tails generated by T4 DNA polymerase (in the absence of dNTP) treatment were unknown [11], [12], and the accurate length of the ss DNA overlaps can be produced using NEases, the comparison of treatments using T4 DNA polymerase and NEases on ss DNA overlaps annealing efficiency was performed (Fig. 3). To study whether gaps generated by an uncontrollable digestion of T4 DNA polymerase (in the absence of dNTP) can result in a decreased efficiency, three groups of assembling studies were conducted (Fig. 3). The first group used T4 DNA polymerase to generate ss DNA overlapping tails; the second one employed nicking endonuclease to form ss DNA overlaps consisting of a two base pairs non-overlap (2 bp gaps); the third one adopted a nicking endonuclease to generate a seamless ss DNA overlaps (Fig. 3). The homologous overlaps from each group were designed with lengths of 5 bp, 10 bp, 15 bp, 20 bp, and 25 bp, respectively. After the T4 DNA polymerase or NEases treatments and further incubation at 37°C, electrophoresis of all mixtures was performed to study the annealing effects (Figs. 3A, 3B and 3C). Results of the chemical transformation showed that the lowest efficiency (Figs. 3A and 3D and Table 4) was from T4 DNA polymerase treatment, which might produce non-complementary ss DNA inside of the ss tail besides the homologous ss overlap, while better results were observed from nicking enzyme treatment with 2 bp gaps (Figs. 3B and 3D and Table 4). On the other hand, NE-LIC without gaps produced the highest efficiency results (Figs. 3C and 3D and Table 4). All phenomena demonstrated that a controllable overhang length enabled higher assembly efficiency.

Fragments of the attL-R6K, attR-kan, and cat encoding chloromycin, respectively, were assembled using 15 bp single-stranded overlaps (Fig. 4). The gel electrophoresis displayed successful assembly of the three fragments. The reaction mixture was plated on Kan^R+Cm^R Petri disks. 248 colonies that were results of the successful 3′-end ss DNA overlaps annealing, were observed on Kan^R+Cm^R Petri disks. In comparison, for 5′-end ss overlaps annealing, 476 colonies were grown on Kan^R+Cm^R Petri disks. All of the colonies on Kan^R+Cm^R Petri disks contained the cat selection marker. The colonies were verified via DNA sequencing, all indicated a correct assembly. In order to investigate the assembly efficiency, the reaction mixture was plated on Kan^R Petri disks at the same time. 287 and 532 colonies were found for the 3′ and 5′-end ss overlaps annealing, respectively.

Poly-3-hydroxybutyrate (PHB) synthesis pathway cloned from Ralstonia eutropha was assembled using 15 bp single-stranded overlaps annealing (Fig. 5). Six DNA fragments including attL-R6K fragment, attR-kan fragment, R. eutropha native promoter (Re promoter), genes of beta-ketothiolase (phbA), acetoacetyl-CoA reductase (phbB) and PHA synthase (phbC), were amplified with specific primers containing 15 bp overlaps. The fragments were digested either by the Nt.BbvCI (3′-end ss DNA overlaps) or by the Nt.BspQI (5′-end ss DNA overlaps). After the denaturation process, the incubation led to formation of the ds linear DNA fragment assembled by the above six fragments. They were transformed into E. coli EC 100D pir-116 harboring pAH83CI.

When assembled all six fragments together in one reaction at 37°C, no colony containing the correct assembled fragments was found. Thus, a different approach was adopted by first incubating two fragments from the six fragments together, such as fragments of the attL-R6K and Re promoter, fragments of genes of of PHB synthase (phbC) and beta-ketothiolase (phbA), and fragments of attR-kan and gene of acetoacetyl-CoA reductase (phbB) (Fig. 5A). Consequently, the three fragment groups each containing two fragments were mixed together and incubated at 37°C for another half an hour. The expected assembly was transformed into the E. coli. Several colonies were observed on the Petri disks after 24 h. PCR verification based on primers phaBRtest/reFtest confirmed all the three randomly picked colonies to have the correct assembly of the six fragments mentioned above (Table 2). As a result, PHB accumulation was detected in the three positive recombinants (Table 5). The three recombinants grew to over 13 g/L cell dry weight (CDW) containing over 40% PHB in their CDW in 48 h of growth, demonstrating the success of the multiple-fragments assembly using NE-LIC.

E. coli strain EC100D pir-116 containing plasmid pAH83CI was used for all the DNA assemblies. Plasmid pUKG was used as a template for attL-R6K and attR-kan fragments. Chloromycetin gene was cloned from pKD3 plasmid. Plasmid pBHR68 containing Ralstonia eutropha PHB synthesis operon was used as a template for cloning Re promoter, phbC, phbA, and phbB which encodes R. eutropha native promoter, PHA synthase, beta-ketothiolase, and acetoacetyl-CoA reductase, respectively. All strains and plasmids used in this study were listed in Table 1. Detailed primer information for fragments was listed in Table 2.

Nicking enzymes (Nb.BbvCI, Nt.BspQI and Nb.BtsI), T4 DNA polymerase and T4 DNA ligase used in this study were purchased from New England Biolabs (Ipswich, MA, USA). DNA fragments were amplified using Fast-pfu DNA polymerase (TRANSGEN, Beijing, China) and purified by the OMEGA E.Z.N.A. Gel Extraction kit (Omega Bio-Tek, USA). The PCR manipulation of DNA fragments were performed as stated below: after an initial 8 minutes denaturation at 98°C, 5 cycles of 30 seconds at 98°C, 30 seconds at 52°C for annealing, and at 72°C for 15–30 seconds for extension (depending on the length of the fragment), followed by another 30 cycles of PCR amplification with higher annealing temperature as 65°C. Finally an extension step at 72°C for 5 minutes allowed completion of this DNA manipulation process.

After the amplification and purification processes, DNA fragments with designed overlaps were digested by 10 U of Nb.BbvCI or Nt.BspQI at 37°C for 1 h. Then the digested fragments were heated to 90°C for 5 min for denaturation, and then chilled to 4°C. The DNA concentrations were determined by Nano Drop Spectrophotometer ND2000 (Thermo Scientific,Wilmington,USA). For NE-LIC coupled with in vivo circularization, an equal molar ratio of the fragments was mixed to anneal at 37°C for 1 h into a linear form and then electroporated into competent cells of E. coli EC100D pir-116 harboring pAH83CI, to allow in vivo circularization. Following electroporation, cells were suspended in LB broth and incubated under the following conditions: 30°C for 30 min; 37°C for 30 min and finally another 30°C for 30 min. The cells were then spread onto 10 µg/mL kanamycin petri disk and incubated at 37°C for 16 h. The schematic diagrams are shown in Figures 1A and 1B. The constructs were verified by PCR and DNA sequencing. The buffer for this study was optimized among buffers of T4 DNA ligase, T4 DNA polymerase and Fast-pfu polymerase.

A fragment with two Nb.BtsI sites of 22 bp separated length in the middle was amplified from plasmid pUC19 using primers BtsIf/BtsIr (Table 2), and further purified using DNA OMEGA E.Z.N.A. Gel Extraction kit (Omega Bio-Tek, USA). It was then digested by Nb.BtsI nicking enzyme, and denatured at 70°C, 75°C, 80°C, 85°C, 90°C or 95°C to test the optimal temperature for splitting into two fragments.

Competing primers that are complementary to the ss DNA released during the denaturation process was added to the denaturation system. Competing primers were designed with the same length of the ss DNA overlaps. The denatured products were purified by OMEGA E.Z.N.A. Gel Extraction kit (Omega Bio-Tek, USA).

The 2-fragment NE-LIC coupled with in vivo circularization of the R6kγ and Kan fragments was used for the optimization of overlapping lengths. The junction of the R6kγ and Kan fragments was designed to generate different overlaps of 5 bp, 10 bp, 15 bp, 20 bp or 25 bp. Annealing with or without T4 DNA ligase was studied. Electrophoresis of annealing products was performed (Fig. 2).

The anneals of R6kγ and Kan fragments with 5 bp, 10 bp, 15 bp, 20 bp and 25 bp complementary overlaps were designed, respectively. For the group with uncontrollable gaps, all the assembled fragments were treated with T4 DNA polymerase at 37°C for 5 min, and then the reaction was terminated using 0.1 vol of 10 mM 2′-deoxycytidine 5′-triphosphate (dCTP). For the group with 2 bp designed gaps, the 2-fragment NE-LIC coupled with in vivo circularization of the R6kγ and Kan fragments was used to study the effect. The 2 bp gaps were designed inside the complementary region close to the inner end. For the group without gap, the junction of the R6kγ and Kan fragments was designed to generate overlaps that anneal seamlessly. Electrophoresis of annealing products was performed (Fig. 3).

E. coli EC100D pir-116 harboring the PHB synthesis operon phaCAB constructed by NE-LIC coupled with in vivo circularization was incubated at 37°C in LB medium containing (g/L) 5 yeast extract, 10 tryptone and 10 NaCl for 12 h at 200 rpm on a rotary shaker (Series 25D, NBS, New Brunswick, USA). Then they were inoculated into the shake flasks placed on the rotary shaker at 200 rpm placed with 500 ml conical flasks containing 50 ml Terrific Broth (TB) medium containing (g/L) 12 tryptone, 24 yeast extract, 9.4 K₂HPO₄, 2.2 KH₂PO₄ and 4 ml/L glycerol supplemented with 20 g/L glucose for 48 h [30]. Additionally, 50 mg/L kanamycin was used for maintaining the stability of the plasmids. PHB analysis method was performed as described by Zhou et al [30].