Compositions and methods for improved in vitro assembly of polynucleotides - Patents.com
Patent Information
- Application Number
- JP2023536435
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-06-23
- Filing Date
- 2021-12-15
- Publication Date
- 2024-12-04
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing DNA assembly methods, such as Golden Gate assembly, face challenges with type IIS restriction endonucleases that create undesirable internal cleavage sites, leading to reduced assembly efficiency and fidelity, especially in multi-fragment assemblies, and T4 DNA ligase biases in ligation, which affect the accuracy of large-scale fragment assembly.
The use of PaqCI, a type IIS restriction endonuclease with a 7-nucleotide recognition sequence and PaqCI activator oligonucleotides, along with optimized ligation conditions and a computer-implemented method to select overhangs, ensures complete cleavage and reduces biases in ligation, enhancing the fidelity and frequency of large DNA molecule assembly.
This approach significantly improves the efficiency and accuracy of multi-fragment DNA assembly, allowing for the creation of large DNA molecules with higher fidelity and frequency, as demonstrated by improved colony formation and sequence fidelity in assembled DNAs.
Smart Images

Figure 00000080_0000 
Figure 00000080_0001 
Figure 00000080_0002
Abstract
Description
[Background technology]
[0001] The ordered in vitro assembly of large DNA molecules from smaller building block DNA molecules is a key feature of synthetic biology. A variety of methods have been developed, including sequence- and ligation-independent cloning (SLIC) (Li et al., Nat. Methods Res vol. 4, 251-256 (2007)), the Golden Gate method (Engler et al., PLOS one 3, e3647 (2007); Engler et al., PlosOne e5553 (2009)), circular polymerase extension cloning (CPEG) (Quan et al., PlosOne 4, e6441 (2009)), NEBuilder® (New England Biolabs, Ipswich, MA), seamless ligation cloning extraction (SLICE) (Zhang et al., NAR 40, e55 (2012)) and ordered gene assembly in Bacillus subtilis (OGAB) (Tsuge et al., Scientific Reports, 5, 10655 (2015)). These methods differ with respect to in vitro and in vivo assembly methods, homologous recombination approaches, and different uses of exonucleases and ligases. The success of any assembly method depends on the frequency and fidelity of assembled fragments, but there has been little systematic analysis of the criteria on which these parameters depend.
[0002] In a method called Golden Gate assembly, fragments of DNA are created using a restriction endonuclease that generates single-stranded overhangs on the double-stranded DNA. Ligation then occurs between the overhangs of multiple different fragments, allowing the fragments to assemble into a single double-stranded molecule. A method for identifying preferred overhangs for polynucleotide fragment assembly under specific criteria for a desired number of fragments is described in WO2020 / 081768. Selection of optimized overhangs using T4 DNA ligase has been based on varying the sequence of the overhangs and collated using computer software based on extensive sequencing of the assembled fragments. Other factors, including incubation temperature and time, are varied and these results are incorporated into the computer analysis. Systematic analysis of the fidelity and efficiency of fragment assembly, as well as the accessibility of the resulting data in a user-friendly format, has been shown to facilitate faithful assembly of multiple fragments in a desired order in a time-efficient manner.
[0003] One aspect of the Golden Gate assembly method is the reliance on type IIS restriction endonucleases and ligases. Type IIS endonucleases that recognize 5 or 6 bases can produce undesired internal cleavage sites. This can be removed by site-directed mutagenesis or by designing assembly junctions within the recognition sequence, but the removal strategies are time consuming and increase costs. Internal sites significantly reduce assembly efficiency, as the completed construct becomes susceptible to digestion by restriction enzymes present in the assembly reaction, leading to incorrect and undesired assemblies. Therefore, it is desirable to use type IIS endonucleases that recognize 7 nucleotides for cleavage. Such enzymes are particularly useful for multi-fragment assemblies where the assembly is complex and maximum efficiency is desired. Furthermore, endonucleases that can cleave completely and have no detectable star activity are preferred.
[0004] Grigaite et al., Nucleic Acid Research 2002, Vol. 30, e123, describes AarI, a type IIS endonuclease with a 7-nucleotide recognition sequence. Unfortunately, this endonuclease has star activity and does not cleave DNA completely. Since neither the DNA or protein sequence of AarI nor the buffer requirements are known, the option of improving this enzyme is unavailable. Therefore, there is a need for further improved restriction endonucleases that recognize 7 nucleotides to generate 4-base overhangs and can cleave completely.
[0005] Another aspect of the Golden Gate assembly method is its reliance on T4 ligase. Bias in the ligation of various complementary overhangs has been detected with T4 DNA ligase (Potapov et al., ACS Synthetic Biology, 7, 2665-2674 (2018); Nilsson et al., Nucleic Acids Res. 10:1425-1437 (1982); Goffin et al., Nucleic Acids Res. 15:8755-8771 (1987); Wu et al., Gene, 76:245-254 (1989); Harada et al., Nucleic Acids Res., 21, 2287-2291 (1993); Showalter et al., Chem Rev. 106:340-360 (2006); Engler et al., PlosOne e5553 (2009); Engler et al., Methods Mol. Biol., 729:167-181 (2011); Engler et al., Methods Mol. Biol., 1116, 119-131 (2014)). This bias, with less than perfect fidelity due to ligation of mismatched sequences in the ligated DNA, becomes a significant problem for large numbers of fragments in ordered assembly reactions. No general rules for selecting overhang sets for large fragment assemblies have been identified. Instead, assembly design may best be achieved for each individual case using computational tools that can compare annealed overhang data to provide optimized overhang sets or evaluate existing overhang sets. As the use of assembly of large sets of fragments increases, so does the need to improve the computational tools available for data-optimized assembly design. [Prior art documents] [Patent documents]
[0006] [Patent Document 1] International Publication No. 2020 / 081768 [Non-patent literature]
[0007] [Non-Patent Document 1] Li et al., Nat. Methods Res 4, 251-256 (2007) [Non-Patent Document 2] Engler et al., PLOS one 3, e3647 (2007) [Non-Patent Document 3] Engler et al., PlosOne e5553 (2009) [Non-Patent Document 4] Quan et al., PlosOne 4, e6441 (2009) [Non-Patent Document 5] Zhang et al., NAR 40, e55 (2012) [Non-Patent Document 6] Tsuge et al., Scientific Reports, 5, 10655 (2015) [Non-Patent Document 7] Grigaite et al., Nucleic Acid Research 2002, Vol. 30, e123 [Non-Patent Document 8] Potapov et al., ACS Synthetic Biology, 7, 2665–2674 (2018) [Non-Patent Document 9] Nilsson et al., Nucleic Acids Res. 10:1425–1437 (1982) [Non-Patent Document 10] Goffin et al., Nucleic Acids Res. 15:8755~8771 (1987) [Non-Patent Document 11] Wu et al., Gene, 76:245–254 (1989) [Non-Patent Document 12] Harada et al., Nucleic Acids Res., 21, 2287-2291 (1993) [Non-Patent Document 13] Showalter et al., Chem Rev. 106:340-360 (2006) [Non-Patent Document 14] Engler et al., Methods Mol. Biol., 729:167–181 (2011) [Non-Patent Document 15] Engler et al., Methods Mol.Biol., 1116, 119–131 (2014) Summary of the Invention [Means for solving the problem]
[0008] Provided are synthetic self-complementary oligonucleotides characterized by a double-stranded region and a single-stranded loop, where the double-stranded region contains a recognition sequence for PaqCI® (New England Biolabs, Inc.), has 3' and 5' ends that cannot be ligated, and cannot be cleaved by PaqCI. PaqCI is defined herein to include variants that have 10% or less amino acid alterations compared to wild type and retain DNA recognition specificity and cleavage properties. The oligonucleotides may be further defined by any one or more of the following features: a double-stranded region having a length of 10-50 base pairs; a length of the oligonucleotide less than 110 nucleotides; a 3' end of the oligonucleotide that is not a 3' hydroxyl; a 5' end of the oligonucleotide that is not a 5' phosphate and / or a recognition sequence that is CACCTGC; and occurs only once in the oligonucleotide.
[0009] A reaction mixture is provided comprising the above synthetic self-complementary oligonucleotide and PaqCI restriction endonuclease or a variant thereof having an amino acid sequence having at least 90% amino acid sequence identity with SEQ ID NO:1, wherein PaqCI is defined herein to include variants having 10% or less amino acid modifications compared to the wild type and that retain DNA recognition specificity and cleavage properties. The reaction mixture features may include one or more of the following features: a ratio of PaqCI to synthetic self-complementary oligonucleotide ranging from 1 unit of PaqCI:0.75 picomoles to 9 picomoles of oligonucleotide; comprising a double-stranded DNA substrate and / or a ligase; the DNA substrate contains one or more recognition sequences for PaqCI and can be cleaved by PaqCI to result in a four-base overhang; the recognition sequence in the DNA substrate is CACCTGC; the DNA ligase is selected from the group consisting of T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, PBCV-1 DNA ligase and hLig3; a ratio of PaqCI to ligase is 2.5 to 20 PaqCI units to 200 to 800 ligase units; the reaction mixture comprises a plurality of plasmids or PCR products, each containing fragments flanked by binding sites for PaqCI, where cleavage of the plasmids or PCR products with PaqCI or a variant thereof results in fragments with different four-base overhangs.
[0010] A method is provided that includes the steps of: (a) obtaining a reaction mixture comprising (i) the synthetic oligonucleotide described above; (ii) PaqCI; (iii) a ligase; and (iv) a library of DNA substrates, each having at least one PaqCI recognition sequence and cleavage site; (b) cleaving the library of DNA substrates with PaqCI to generate fragments having four-base overhangs; and (c) ligating complementary four-base overhangs to each other to produce an ordered assembly of fragments.
[0011] The method may include the following features: the DNA substrates in the library are selected from one or more of the group consisting of PCR products, plasmids, genomes, or chromosomes; step (c) may further comprise ligating the ordered assembly in a destination vector or a viral genome; the destination vector is a plasmid or a chromosome; the ligase is selected from the group consisting of T4 DNA ligase, T3 DNA ligase, T7 The enzyme may be selected from the group consisting of DNA ligase, PBCV-1 and Hitrigase 3; there are 10-100 DNA substrates having unique sequences and the ordered assembly comprises 10-100 fragments that are ligated to each other in step (c); at least 20 DNA substrates having unique sequences are included in the reaction mixture and the ordered assembly comprises at least 20 fragments that are ligated to each other in step (c); and the reaction mixture may further comprise a DNA repair enzyme, e.g., EndoMS, a deadenylase, e.g., yeast deadenylase, and / or a crowding agent, e.g., polyethylene glycol (PEG) having a molecular weight in the range of 600-8000.
[0012] The method may further comprise, in step (a), (i) the computer tool generating from the dataset an optimized fidelity and / or frequency score for a set of four-base overhangs for the library of DNA, the optimized fidelity and / or frequency score being derived from data on annealing of complementary sequences and data derived from ligase activity against different four-base overhangs; and / or (ii) the computer tool identifying a set of four-base overhangs for the reaction mixture using the computer tool that provides cut points in the in silico sequence to generate fragment sequences for participating in ordered assembly via the optimized four-base overhangs.
[0013] A kit is provided that contains the synthetic self-complementary oligonucleotides described above and PaqCI including variants as defined above. Additional components of the kit may include one or more of the following: a ligase; a repair enzyme that is a mismatch-specific endonuclease such as EndoMS, a cofactor selected from the group consisting of a deadenylase and a crowding agent such as polyethylene glycol (PEG) having a molecular weight in the range of 600-8000; and instructions for use to synthesize large DNA from component fragments with 4-base overhangs. The reagents in the kit may be combined or in two or more containers. In one embodiment, at least one of the oligonucleotides, the ligase and the PaqCI variant is lyophilized or immobilized on a solid substrate such as a two- or three-dimensional surface.
[0014] 1. A computer-implemented method for selecting a set of overhangs for an ordered assembly reaction performed under selected ligation conditions, comprising: (a) receiving (i) a desired number of overhangs for the assembly reaction and (ii) an overhang length; (b) selecting a set of overhangs from an overhang table, where the selected set of overhangs has the desired number of overhangs received in (i) and the overhang length received in (ii); (c) selecting a ligase from a plurality of different ligases for ligating the overhangs with reduced bias; (d) calculating a ligation fidelity score of the selected ligase for each individual overhang in the set, where each The method includes the steps of: (a) calculating an overall ligation fidelity score for the set of overhangs based on the calculated ligation fidelity scores for each of the individual overhangs output in step (d); (b) repeating steps (e) to (e) until a plurality of overall ligation fidelity scores have been calculated, each for a different set of overhangs; and (c) providing a set of overhangs having a suitable overall ligation fidelity score for a selected ligase.
[0015] One or more features of the computer-implemented method include: each individual overhang in the set of overhangs selected in step (b) is unique within the set, is not complementary to another overhang in the set, and is not palindromic; calculating the ligation fidelity score in step (c) further comprises referencing ligation frequency tables and bias tables for different ligases that contain individual experimentally defined measures of the number of ligation events and / or mismatch events; between each individual overhang and all overhangs in the set and their complements, and between the complements of each individual overhang and all overhangs in the set and their complements. calculating the number of ligation and / or mismatch events that occur between each individual overhang and its complement compared to the total number of ligation events that occur in the set of overhangs; the set of overhangs corresponds to individual overhangs on each end of a plurality of double-stranded polynucleotide fragments of an ordered assembly in the target polynucleotide, each individual overhang being a single-stranded sequence of 2-5 nucleotides such that each end of each polynucleotide fragment has a different overhang, and the ordered assembly of the fragments is the product of annealing of an overhang on one end of a polynucleotide with a complementary overhang on one end of an adjacent polynucleotide.
[0016] Another feature of the method may include, in step (a), receiving (iv) a nucleotide sequence of the assembly; and (v) a set of intervals at which the nucleotide sequence of (iv) can be enzymatically cleaved, and identifying a non-redundant set of subsequences with intervals that are the same length as the length of the overhangs input in (ii), each subsequence having an overhang; the method further includes (h) saving the non-redundant set of subsequences with a set of overhangs that have a preferred overall fidelity score.
[0017] Another feature may include defining each interval in (v) by start and end coordinates in a nucleotide sequence of the assembly.
[0018] Another feature may include, in step (e), repeating steps (b) to (d) at least 1000 times.
[0019] Another feature may include in step (a) receiving selected experimental conditions for enzymatic cleavage and ligation for the ordered assembly of polynucleotide fragments.
[0020] Another feature may include receiving selected experimental conditions to provide a set of overhangs in step (g) that have suitable fidelity and / or frequency scores for annealing and ligation with the selected ligase.
[0021] Another aspect may include running the computer-implemented method described above and receiving an output containing the set of overhangs identified in step (g) and / or, where (iv) and (v) are inputs, receiving sequences of a set of polynucleotide fragments of the ordered assembly, the ends of the fragments being defined by the overhangs identified in step (g).
[0022] The computer-implemented method may include obtaining sequences of a set of polynucleotide fragments having an identified non-redundant set of subsequences at intervals that can be enzymatically cleaved to produce the identified overhangs. Another feature of the method may include establishing that the selected experimental conditions and computer-generated set of overhangs are suitable for ordered assembly of the set of selected polynucleotide fragments with an effective amount of fidelity and frequency of complementary annealing and ligase-dependent ligation relative to the number of fragments in the set. The experimental conditions (ligation conditions) are, for example, those in which the ligase is a wild-type T4 DNA ligase or a variant thereof selected from thermostable T4 DNA ligase and salt-tolerant T4 DNA ligase, and the ligase is a thermostable variant of any of T4 DNA ligase, T7 DNA ligase, hLig3 DNA ligase, T3 DNA ligase, PBCV-1 DNA ligase, T4 DNA ligase, T7 DNA ligase, hLig3 DNA ligase, T3 DNA ligase, and PBCV-1 DNA ligase, and those in which the ligase is a thermostable variant of any of T4 DNA ligase, T7 DNA ligase, hLig3 DNA ligase, T3 DNA ligase, and PBCV-1 DNA ligase. selecting a DNA ligase having suitable fidelity and frequency scores for ligating a set of polynucleotide fragments containing four-base overhangs, selected from the group consisting of any high salt stable variant of DNA ligase; and / or selecting a Type IIS restriction endonuclease having a unique DNA recognition sequence for enzymatic cleavage of a set of polynucleotide fragments containing recognition sequences such that each polynucleotide fragment so cleaved contains an overhang sequence derived from the set of overhang sequences, e.g., a Type IIS restriction endonuclease having a seven-base recognition sequence, e.g., PaqCI.
[0023] Other examples of selected experimental conditions for the ordered assembly of a target polynucleotide from a set of polynucleotide fragments include ligation conditions including one or more of salt concentration, DNA repair enzyme, temperature range and / or thermal cycling conditions for cleavage and ligation. For example, the salt concentration ranges from 50 mM to 150 mM salt, the DNA repair enzyme is EndoMS or T7 EndoI, the temperature range is from 37°C to 50°C, and the thermal cycling conditions are selected from drop-down, touch-down and touch-up temperature cycling.
[0024] In a further embodiment of the method, further features may include a nucleotide sequence of the assembly selected from a viral genome, a prokaryotic genome, an operon, and a metabolic pathway, and the number of polynucleotide fragments for producing the assembly ranges from 2 to 100 fragments.
[0025] Computer readable media is provided for carrying out the described methods with suitable software.
[0026] Methods for synthesizing a target polynucleotide are provided, comprising: (a) using a computer-implemented method to obtain a set of overhangs with a favorable overall fidelity score under a set of experimental conditions including a selection of a ligase, wherein the computer instructs an automated instrument or a user to assemble a set of enzymatically obtained or chemically synthesized polynucleotide fragments having sequences, optionally determined by a computer or a user, under a selected set of experimental conditions, at least in part determined by a user; (b) combining a ligase, a restriction endonuclease and the polynucleotide fragments under selected experimental conditions in an instrument or in a reaction tube, thereby enabling an optionally automated ordered assembly of the target polynucleotide; and (c) optionally introducing the target polynucleotide into (i) a bacterial cell or (ii) an in vitro system for expression of a gene or genes.
[0027] The method allows for the assembly of a target polynucleotide by repeating steps (a) and (b), where in a first round, the polynucleotide fragments are less than 1000 bases long, the assembled fragments form said tentative target polynucleotide, and the tentative target polynucleotide forms polynucleotide fragments for the next round of ordered assembly to form a final target polynucleotide. The set of polynucleotide fragments in step (a) is 2-100 fragments, more specifically 20-100 fragments or at least 20 fragments. The method may further comprise a step of performing multiplex amplification of the set of polynucleotide fragments prior to step (b). In the method, the target polynucleotide may be DNA that can be transcribed to form a target RNA. The target polynucleotide may be DNA, which is expressed in a cell to produce one or more proteins. For example, the target protein may be part or all of a metabolic pathway, a viral genome, or an immune cell gene.
[0028] Provided is a method for performing ordered DNA assembly from 20-100 DNA fragments to generate a large DNA, the method comprising the following steps, which may be performed in any order: (a) obtaining instructions from a computer design tool for an optimal set of four-base overhang sequences for ligating the 20-100 fragments in an ordered assembly reaction, where the computer design tool calculates the optimal set of overhangs from one or more sets of data, each set of data obtained from a frequency and fidelity analysis of individual ligase preferences for all combinations of four-base overhangs; and (b) obtaining 20-100 fragments with an optimal set of four-base overhangs for ligation with a selected ligase in an ordered assembly reaction to generate the large DNA.
[0029] The method may include adding a Type IIS restriction endonuclease recognition sequence to the 20-100 fragments using polymerase chain reaction (PCR), or inserting the 20-100 fragments into 20-100 plasmids having a Type IIS restriction endonuclease recognition sequence at an insertion site at each end of the fragment, or synthesizing the 20-100 fragments with optimized 4 base overhangs.
[0030] A Type IIS restriction endonuclease can be selected that has a recognition sequence of 5'CACCTGC3' and cleavage sites for generating an optimized set of four base overhangs are 5'CACCTGC(N4)3' (SEQ ID NO:2) and 3'GTGGACG(N8)5' (SEQ ID NO:3). For example, the restriction endonuclease is PaqCI. In other aspects, the ligase can be selected from one or more of T4 DNA ligase, T7 DNA ligase, hLig3 DNA ligase, T3 DNA ligase, or PBCV-1 DNA ligase.
[0031] Provided is a method for highly efficient assembly of customized T cells, comprising: (a) identifying a surface antigen on a tumor cell derived from a patient, where the protein is specific to the tumor cell; (b) collecting T cells from the patient; (c) inducing an ordered assembly of DNA fragments with four-base overhangs to form a large DNA that encodes a chimeric antigen receptor that is tumor antigen-specific; and (d) introducing the large DNA into the genome of the T cell that is cleaved by a site-specific CrispR.
[0032] The large DNA in step (d) may be a product of ordered assembly of a plurality of conserved DNA fragments and a plurality of variable DNA fragment sequences, such that at least the conserved DNA fragments are individually stored in a plasmid in a bacterial cell for highly efficient assembly of customized T cells.
[0033] A method is provided for generating a viral genome with mutations, comprising: (a) generating multiple fragments for ordered assembly in the viral genome; (b) selecting 4-base overhangs that allow ligation of multiple mismatches by ligase; and (c) testing the product viral genome for antibiotic activity or as a substrate for vaccine production. An example of this method is when the ligase is a relatively low fidelity ligase, such as hLig3. The method can be accomplished in a highly efficient workflow using a microfluidic or robotic device to handle multiple samples in repeated cycles of ligating fragments to generate DNA of any size from small fragments of DNA. [Brief description of the drawings]
[0034] [Figure 1A] Figure 1A-B show the PaqCI performance comparison (efficiency) for 24-fragment assembly as determined by the number of colonies with a blue phenotype indicative of correct LacL / lLacZ assembly per 50 ul of growth (1 / 20 of total growth). PaqCI from Paucibacter aquatile, together with synthetic activator oligonucleotides (also referred to as "activators" or "oligonucleotides") and T4 DNA ligase, significantly improved the efficiency of colony formation and sequence fidelity in large DNA assembled from 24 fragments compared to AarI. Figure 1A: PaqCI yielded colonies with correct assembly in greater numbers than those observed with AarI over 30 cycles. 1 μg of lambda DNA was digested with either 8 units of PaqCI (New England Biolabs, Ipswich, MA) or AarI (Thermo Fisher Scientific, Waltham, MA) following the manufacturer's recommended protocol. The digests were analyzed on a 1% agarose gel. [Figure 1B]Figure 1A-B show the PaqCI performance comparison (efficiency) for 24-fragment assembly as determined by the number of colonies with a blue phenotype indicative of correct LacL / lLacZ assembly per 50 ul of growth (1 / 20 of total growth). PaqCI from Paucibacter aquatile, together with synthetic activator oligonucleotides (also referred to as "activators" or "oligonucleotides") and T4 DNA ligase, significantly improved the efficiency of colony formation and sequence fidelity in large DNA assembled from 24 fragments compared to AarI. Figure 1B: PaqCI resulted in at least 15% greater fidelity than Aar1 over 30 cycles (5 min 37°C to 5 min 16°C) for the assembly reaction of the 24 fragments of the LacL / LacZ cassette as determined by blue colonies. One microgram of lambda DNA was digested with 8 units of either PaqCI (New England Biolabs, Ipswich, Mass.) or AarI (Thermo Fisher Scientific, Waltham, Mass.) according to the manufacturer's recommended protocol. Digests were analyzed on 1% agarose gels. [Figure 1C] Figure 1C shows the recognition sequence and asymmetric cleavage site for PaqCI to generate a 4-base overhang. One microgram of lambda DNA was digested with either 8 units of PaqCI (New England Biolabs, Ipswich, Mass.) or AarI (Thermo Fisher Scientific, Waltham, Mass.) according to the manufacturer's recommended protocol. Digests were analyzed on a 1% agarose gel. [Figure 1D] Figure ID shows that PaqCI, unlike AarI, cleaves completely and does not exhibit star activity. 1 μg of lambda DNA was digested with either 8 units of PaqCI (New England Biolabs, Ipswich, MA) or AarI (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer's recommended protocol. Digests were analyzed on a 1% agarose gel. [Figure 2A] Figures 2A-C provide a schematic of an assay to determine how fragments with different sequence overhangs are affected by ligation bias and the fidelity of the ligation event. Figure 2A: A library containing randomized 4-base overhangs was synthesized. The randomized overhang pairs of the samples are represented diagrammatically. [Figure 2B] Figures 2A-C provide a schematic of an assay to determine how fragments with different sequence overhangs are affected by ligation bias and the fidelity of the ligation event. Figure 2B: Ligation substrates are ligated with a specific DNA ligase to form correct (same overhang shading) and mismatch-containing (different overhang shading) products. Correct or mismatched products were analyzed using SMRT® sequencing (Pacific Biosciences, Menlo Park, CA). [Figure 2C] Figures 2A-C provide a schematic of an assay to determine how fragments with different sequence overhangs are affected by ligation bias and the fidelity of the ligation event. Figure 2C: Ligation fidelity is defined as the percentage of correct ligations. Ligation bias is detected by differences in the total number of ligation products formed per overhang. [Figure 3Ai]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Ai and 3Aii are T4 DNA ligase. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Aii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Ai and 3Aii are T4 DNA ligase. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Bi]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Bi and 3Bii are T7 ligase. Many 4-nt overhangs are poorly represented. [Figure 3Bii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Bi and 3Bii are T7 ligase. Many 4-nt overhangs are poorly represented. [Figure 3Ci]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Ci and 3Cii are Hiligase 3 (hLig3). CAAG, CCAG, CGAG, CTAG, TAAA, TCAA, TGAA and TTAA are under represented. [Figure 3Cii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Ci and 3Cii are Hiligase 3 (hLig3). CAAG, CCAG, CGAG, CTAG, TAAA, TCAA, TGAA and TTAA are under represented. [Figure 3Di]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Di and Figure 3Dii are T3 ligase. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Dii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Di and Figure 3Dii are T3 ligase. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Ei]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Ei and Figure 3Fii are PBVC-1 ligase. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Eii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Ei and Figure 3Fii are PBVC-1 ligase. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Fi]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Fi and Figure 3Fii are T4 ligase + PEG. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Fii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Fi and Figure 3Fii are T4 ligase + PEG. TAAA, TCAA, TGAA and TTAA are under-represented. [Figure 3Gi]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during a 1 h incubation at 25 °C and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii shows stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (colors represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Gi and 3Gii are T7 ligase + PEG, showing the beneficial effect on ligation using T7 DNA ligase. [Figure 3Gii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during a 1 h incubation at 25 °C and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii shows stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (colors represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figures 3Gi and 3Gii are T7 ligase + PEG, showing the beneficial effect on ligation using T7 DNA ligase. [Figure 3Hi]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Hi and Figure 3Hii are hlig3+PEG. CAAG, CCAG, CGAG, CTAG, TAAA, TCAA, TGAA and TTAA are under represented. [Figure 3Hii]Figures 3Ai / 3Aii-3Hi / 3Hii) show significant variation between different ligases in sequence preferences, with variation observed between correct and mismatch ligations. The number and type of under-represented 4-base sequences also differ between ligases. This reveals at least two types of bias: bias for or against ligation events for specific 4-base overhangs, and bias for ligation of mismatches and against perfect matches and vice versa. 3Ai-3Hi show ligation frequency heatmap matrix (logarithmic scale) of all ligation events. Overhangs are listed alphabetically from left to right (AAAA, AAAC, AAAG...TTTG, TTTT) and bottom to top, with Watson-Crick pairings shown on the diagonal. The matrix shows the ligation frequency for each of the 256 x 4 base overhangs on the X-axis against the 256 x 4 base overhangs on the Y-axis. Each base of the 4-nt overhang is color-coded: T in red, C in blue, G in yellow, and A in green (colors represented in different shades of grey). 100 nM of multiplexed 4-nt overhang substrates were ligated with 1.75 μM T4 DNA ligase in standard ligation buffer during 1 h incubation at 25° C. and sequenced by SMRT sequencing. Overhang sequences were normalized to 100,000 ligation events. 3Ai-3Hii show stacked bar plots of the frequency of ligation products containing each overhang, corresponding to each column of the heatmap in (A). Ligation results of perfect Watson-Crick pairs are shown in blue, and ligation products containing one or more mismatches are in orange (represented in two shades of grey). Certain overhangs are poorly represented, as indicated by the arrows. Figure 3Hi and Figure 3Hii are hlig3+PEG. CAAG, CCAG, CGAG, CTAG, TAAA, TCAA, TGAA and TTAA are under represented. [Figure 4A]FIG. 4A shows how the median ligation and spread of bias as a function of the base content of the overhang vary for different ligases, as determined by the frequency of ligation for each combination of 256 different overhangs from the sequencing library for each ligase. T4 DNA ligase, T7 DNA ligase, T3 DNA ligase, and PBCV-1 DNA ligase have similar median bias (indicated by the black horizontal line) and a similar distribution of positive bias for GC-rich overhangs, but some variation in the amount and degree of negative bias for AT-rich overhangs. T7 ligase has a larger median ligation bias than the other ligases, with very few overhangs ligated very efficiently, and the majority of overhangs ligated much less efficiently. Here, the frequency of ligation (y-axis) is a measure of the efficiency of ligation. Although not shown here by color, each dot was colored differently for 0%, 25%, 50%, and 75% and 100% according to its percentage of GC content. The distribution of dots indicates that GC-rich overhangs tend to ligate more efficiently compared to AT-rich overhangs. The results shown were generated by SMRT sequencing of ligation reactions using 100 nM multiplexed 4-base overhang substrate with 1.75 μM T4 DNA ligase, T7 DNA ligase, human DNA ligase 3, T3 DNA ligase, or PBCV-1 DNA ligase incubated for 1 hour at 25° C. in standard ligation buffer. [Figure 4B]Figure 4B shows that the median ligation fidelity (see black line) of T4 DNA ligase, T7 DNA ligase, human DNA ligase 3, T3 DNA ligase, and PBCV-1 varies, and the spread of GC-rich and AT-rich overhangs across the ligation fidelity profile also varies between enzymes. T7 DNA ligase shows the highest ligation fidelity. hLig3 shows the lowest ligation fidelity and the widest spread of data points below the median line. Ligation fidelity was calculated and plotted for all ligases tested. Ligation fidelity is defined as the percentage of correct (Watson-Crick) ligation events versus incorrect (mismatch) ligation events. [Diagram 5] Figures 5A-5F show that polyethylene glycol (PEG) has a significantly positive effect on ligation frequency for overhangs with relatively low GC (a group of overhangs that generally exhibit lower ligation frequency in the absence of PEG), but a slightly negative effect on ligation fidelity. Ligation frequency and ligation fidelity of overhangs are grouped by GC content. Median values are indicated by horizontal lines (dotted for ligation reactions completed in PEG-free buffer, black for ligation reactions completed in PEG-containing buffer). Figure 5A is T4 DNA ligase (frequency) + / - PEG. Figure 5B is T7 DNA ligase (frequency) + / - PEG. Figure 5C is hLig3 DNA ligase (frequency) + / - PEG. Figure 5D is T4 DNA ligase (fidelity) + / - PEG. Figure 5E is T7 DNA ligase (fidelity) + / - PEG. FIG. 5F is hLig3 DNA ligase (fidelity) + / - PEG. [Figure 6]FIG. 6 illustrates, in a high-level block diagram, a system for generating estimated overall ligation fidelity for a user-specified set of overhang sequences and experimental conditions for achieving a desired outcome. The system utilizes a client 802 having bidirectional data communication 803 with a server 804, which in turn can access storage 806 via 808, where 806 contains a database of 4-base 5'-3' ligation fidelity. This can also be a 2-base, 3-base, or 5-base database. The bidirectional data communication 803 may be implemented using a local connector, such as a local area network (LAN) or wide area network. The server 804 may be a dedicated resident server or may be implemented in the cloud. The data storage 806 may be on-board with the server 804. A user inputs data into the client 802, which may include a browser interface. In addition to running a browser, the client 802 may host a graphical user interface for use in inputting a set of 5'-3' 4 base classical form AGCT or other overhang sets, or for selecting experimental conditions for ligation, such as a selected restriction endonuclease, a selected ligase, a buffer containing PEG, the temperature and time of the reaction, and other experimental details. [Figure 7]FIG. 7 shows the input and output steps of a high level flow diagram for performing an assembly reaction using the system outlined in FIG. 8. The user inputs a set of overhang sequences of any desired length. For example, the set of overhang sequences is 902, which is a set of 5'-3' 4-base overhang sequences. The set includes two or more 4-base overhang sequences such that each member of the set is different from every other member of the set. Each overhang sequence represents a member of a single overhang pair that the user preferably desires to use in an experiment to sequentially ligate multiple double-stranded nucleic acid fragments. The 4-base overhangs may be represented as Watson-Crick overhang pairs. A single overhang pair in a set may vary in terms of ligation fidelity depending on whether a particular sequence is a 5' sequence or its complement. Each member pair is considered separately from other member pairs in the set. The user may select experimental conditions for ligation of fragments with overhangs corresponding to the sequences input in 902. These experimental conditions include, for example, incubation time with the ligase, incubation temperature, and ligation frequency and fidelity of the selected ligase 904. By accessing a database of ligation fidelity for individual overhangs or overhang pairs, the system generates an output describing the ligation fidelity for the input overall overhang sequence set and / or individual overhang pairs within the set 906. The system may further output a graphical matrix display of the ligation fidelity for the selected overhang sequence pairs. If the identified fidelity efficiency of the set of four-base overhang sequences input by the user is rejected by the user, the user can evaluate the ligation of the same set of four-base overhangs under different selected experimental conditions, or input a modified set of four-base overhang sequences under the same or different experimental conditions to determine how to ligate a set of double-stranded nucleic acid fragments in an ordered assembly. [Figure 8]8 is a high level flow diagram showing inputs to add to the system output steps. Each example shows user entry (1302-1306) of inputs that generate outputs 1308-1312. However, the input parameters in 1302-1306 may include: (a) a target nucleic acid sequence, (b) a set of polynucleotide fragments, (c) a partial set of polynucleotide fragments provided by the user and the partial set output by the system, (d) a set of vectors with specific junction sequences, (e) preferred overhang lengths, (f) excluded overhangs, (g) a set of overhangs of specified lengths, (h) a partial set of overhangs of specified lengths provided by the user and the partial set of overhangs output by the system, (i) a ligase, e.g., T4 DNA ligase, T7 DNA ligase, PBCV-1, T3 ligase, hLig3, or Taq (j) any other ATP-dependent DNA ligase or NAD+-dependent ligase, such as a DNA ligase; (j) a restriction endonuclease, such as Esp3I, SapI, BbsI-HF; BspQI, HgaI, BsaBI, BsaJI, BsaI, BsaI-HFv2, BsiI, BsmAI, BsmBI, BsmFI, BsmI, Bsr (k) a preferred number of fragments; (l) a preferred incubation time; e.g., 1 hour, 5 hours, overnight; (m) a preferred incubation temperature; e.g., 37° C. or 45° C.; (n) preferred reaction conditions in addition to time and temperature; (o) a preferred overall ligation efficiency; (p) a minimum acceptable ligation efficiency; (q) a maximum GC content; (r) a maximum AT content; (s) a bias allowed. The output may be replaced or supplemented by any one or more of the following: (a) any of the above parameters not specified by the user; (b) a warning if the ligation fidelity determined from the user input parameters is too low based on manual input of certain parameters.Warnings may include internal RE sites, low efficiency junction sites formed from individual overhang pairs, excessive bias of some ligation events over others; (c) a matrix or tabular format for all identified overhang pairs showing strongly ligated Watson-Crick pairs and mismatched overhang pairs with color-coded occurrence frequency; (d) a ligase fidelity viewer grid showing predicted fidelity of the selected set; (e) provision of alternative solutions with modified specific experimental parameters; (f) a graphical display of the assembly design, annotated if full sequences are used; (g) one or more of exportable fragment / primer sequences for synthesizers. The system then calculates any of the above parameters not provided by the user and uses a database of n-base overhangs and 5'-3' ligation efficiencies under various reaction parameters to achieve the requested output, allowing the user to create an ordered assembly of X fragments with a defined ligation efficiency. [Figure 9A] Figures 9A-9D show how the data presented in Figures 3Ai / 3Aii-3Hi / 3Hii) through 5A-5F can be integrated into a computerized version of the Ligase Fidelity Viewer in WO2020 / 081768, including drop-down menus for overhang length and for user input of predefined selection of overhangs for evaluation of frequency and fidelity parameters, plus additional parameters include the use of PEG and / or aprataxin in buffer selection. [Figure 9B]9A-9D show how the data presented in Figures 3Ai / 3Aii-3Hi / 3Hii) through Figures 5A-5F can be integrated into a computer. In addition to the Ligase Fidelity Viewer in WO2020 / 081768, and the version including drop-down menus for overhang length and for user input of a predefined selection of overhangs for evaluation of frequency and fidelity parameters, additional parameters include the use of PEG and / or aprataxin in buffer selection. A further drop-down menu to add to the user interface page of Figure 9B (Ligase Fidelity Viewer) is a drop-down menu allowing the selection of a ligase, which influences the selection of overhangs based on the frequency, bias, and fidelity data described herein. [Figure 9C] 9A-9D show how the data presented in Figs. 3Ai / 3Aii-3Hi / 3Hii) to 5A-5F can be integrated into a computer. In addition to the Ligase Fidelity Viewer in WO2020 / 081768 and a version including drop-down menus for overhang length and for user input of a given selection of overhangs for evaluation of frequency and fidelity parameters, additional parameters include the use of PEG and / or aprataxin in buffer selection. Fig. 9C shows the GetSet drop-down menu. GetSet is an interface that informs the user how well a selected set of overhangs performs in a specified ligation assembly reaction and whether a particular overhang should be included or excluded from the set. [Figure 9D]Figures 9A-9D show how the data presented in Figures 3Ai / 3Aii-3Hi / 3Hii) to 5A-5F can be integrated into a computer. In addition to the Ligase Fidelity Viewer in WO2020 / 081768, and the version including drop-down menus for overhang length and for user input of a predefined selection of overhangs for evaluation of frequency and fidelity parameters, additional parameters include the use of PEG and / or aprataxin in buffer selection. Figure 9D shows a drop-down menu of SPLITSET, which informs the viewer which sites should be included and which sites should be excluded in the in silico sequence for generation of fragments from the corresponding DNA by targeted cleavage or DNA synthesis. [Figure 10] Figure 10 shows that PEG increases the frequency of colonies obtained from multi-fragment assembly using T4 DNA ligase and BbsI-HF restriction endonuclease for a particular concentration of DNA compared to the same DNA in the absence of PEG. All PEG sizes showed some improvement. Preferred embodiments included PEG3350 and PEG6000. [Figure 11] FIG. 11 shows that PEG6000 allows the use of 10-fold less DNA to achieve substantial colony display after assembly of 24 pieces of DNA using T4 DNA ligase and BbsI-HF. [Figure 12A] Figures 12A and 12B show that 50 DNA fragments with overhangs determined by the computational tool described in Figures 9A-9D, including adjustment of the ligation priority of T4 DNA ligase, allowed improved efficiency of assembly of a T7 viral genome from the 50 fragments, as determined by plaques on a bacterial lawn. [Figure 12B]Figures 12A and 12B show that 50 DNA fragments with overhangs determined by the computational tool described in Figures 9A-9D, including adjustment of the ligation priority of T4 DNA ligase, allowed improved efficiency of assembly of a T7 viral genome from the 50 fragments, as determined by plaques on a bacterial lawn. [Figure 13A] Figures 13A and 13B show that the plaques obtained on the bacterial lawn indeed contain intact phage T7 DNA. [Figure 13B] Figures 13A and 13B show that the plaques obtained on the bacterial lawn indeed contain intact phage T7 DNA. [Figure 14] Figure 14 shows that the percentage of colonies containing correctly assembled constructs was maintained at least 50% for 52 fragments using the tools described herein to design overhangs for correct end joining. These results are obtained from a one-pot fragment assembly reaction. [Figure 15] Figure 15 shows a schematic diagram of how the improved multi-fragment assembly method can be used for the scale-up of Car-T cell therapy for thousands of individual patients. Tumor cells from individual patients are analyzed to find the DNA sequence of genes encoding their unique tumor-specific antigens and neoantigens including these mutations. The patient's own T cells are removed and engineered to insert the assembled genes into the target sites in the genome recognized and cleaved by CRispR. The T cells can then be reintroduced into the patient to destroy the tumor cells. Here, a subset of the multiple components required to synthesize the tumor antigen is preserved and a subset of the components is not preserved. The entire region of interest may be maintained in a ready-to-use plasmid library, and individual non-conserved fragments with identified mutations may be used in the assembly reaction. In this way, the entire gene does not need to be created de novo for each patient, allowing a higher throughput of samples in the workflow. [Figure 16]Figure 16 shows a schematic of phage engineering to treat drug-resistant bacterial infections for potential antibiotic solutions, where the phage genome is split into pieces and various mutations are introduced into any one or more of the pieces. Once assembled using the multi-fragment system described herein, the engineered phages can be assayed for their ability to invade and destroy target bacteria. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs. Singleton et al., DICTIONARY OF MICROBIOLOGY AND MOLECULAR BIOLOGY, 2nd ed., John Wiley and Sons, New York (1994) and Hale&Markham, THE HARPER COLLINS DICTIONARY OF BIOLOGY, Harper Perennial, NY (1991) provide those skilled in the art with the general meaning of many molecular biology-related terms used herein. Certain terms are used herein, and explanations for them are provided below.
[0036] As used herein, the term "target polynucleotide" (or "large DNA") refers to the end product of ligation-based ordered assembly of fragments, which can be DNA, RNA, or a mixture thereof.
[0037] As used herein, the term "polynucleotide fragments" (or "DNA fragments") refers to building blocks that, when assembled, generate a target polynucleotide. These building blocks may be derived from sequence databases and may contain promoter sequences, enhancer sequences, coding sequences, etc. Polynucleotide fragments may be produced by chemical synthesis (IDT, Coralville, IA) or by enzymatic synthesis, for example using terminal transferase-based synthesis. Fragments thus produced may be assembled in a preliminary step from the products of chemical and / or enzymatic synthesis to form larger polynucleotide fragments suitable for assembly into genes. When individual polynucleotide fragments are amplified from a template, for example by PCR, their length does not exceed the processivity of the polymerase used for amplification. For example, in the case of Taq polymerase, amplicons rarely exceed 5 kb to 10 kb and may have a minimum length of 15 nucleotides long.
[0038] As used herein, the term "oligonucleotide" refers to a multimer of at least 10, e.g., at least 15 or at least 30 nucleotides in the intended context. In some embodiments, oligonucleotides may range from 15 to 500 nucleotides in length or more. Any oligonucleotide as used herein may be composed of G, A, T, and C, or bases that can reliably base pair with complementary nucleotides.
[0039] As used herein, the term "sequence" refers to a series of letters where each letter corresponds to a base, e.g., A=adenine, T=thymine, U=uracil, G=guanine, and C=cytosine. Modified nucleotides (nucleosides (bases and deoxyribose) and phosphates) may optionally be included in the overhang sequence and thus in the ligation efficiency database. Multiple letters in a particular order symbolically describe the base (also called nucleotide) composition of the molecule.
[0040] As used herein, the terms "perfect match," "complementary," and Watson-Crick pairing refer to the hydrogen-bonded pairing of bases on separate strands of double-stranded DNA, where A matches T or U, and G matches C, respectively.
[0041] As used herein, the term "junction" refers to the position in the target polynucleotide where the component polynucleotide fragments are joined by ligase. The term "junction" also refers to the position in the sequence of the target polynucleotide in the database where fragmentation is recommended for the assembly of the target polynucleotide from the optimized fragment set. The context of the phrase "junction" makes it clear which of the two meanings is intended. The assembly method described herein may be used to create an intact junction in the target polynucleotide. An intact junction in the target polynucleotide means that the junction in the target polynucleotide is indistinguishable from the corresponding position in the original polynucleotide sequence.
[0042] As used herein, the term "overhang" refers to a single-stranded region at the end of a double-stranded fragment polynucleotide, e.g., DNA. Overhangs are preferably formed by an enzyme that generates staggered cleavage of the nucleic acid on both strands of the duplex outside of the recognition region. Overhangs are usually 5' overhangs. Overhangs can be defined by their length and their sequence. For example, there are 256 different possible 4-base overhangs (4 4 ) are included. 2-, 3-, 4-, and 5-base overhangs generated by restriction endonuclease cleavage are exemplified herein. Overhangs can contain from 2 to 8 bases, although 3- or 4-base overhangs are generally preferred. The preference comes from the availability of restriction endonucleases that cleave double-stranded DNA outside of the recognition site to generate 3- or 4-base overhangs, and the number of possible overhang pairs in the set sufficient to optimize ligation of multiple polynucleotide fragments to form a target polynucleotide.
[0043] Matching an overhang from one polynucleotide fragment with a second complementary overhang of a second polynucleotide fragment results in a junction when a ligase is added to the mixture and ligation occurs depending on the preference of the ligase for the overhang sequence and its complement. The first overhang and the second complementary overhang are referred to as an overhang pair or complementary overhang. Without wishing to be limited by theory, it is proposed herein that combining a ligase with a restriction endonuclease in a single assembly reaction mixture results in a significant reduction in improper hybridization and ligation events. These improper events occur when the cleavage products consisting of the overhang and the restriction endonuclease recognition sequence recombine with the cleaved assembly fragment or recombine with another cleavage product. In both scenarios, the sequence is cleaved again by the restriction endonuclease in the reaction mixture, liberating the polynucleotide overhang for proper ligation to a matching polynucleotide fragment partner. Another mismatch event can occur when non-complementary overhangs anneal, resulting in mismatches, which generally only result in one or two mismatches and can affect the order of assembly unless the occurrence of mismatches in the annealed overhangs is taken into account in the assembly strategy.
[0044] As used herein, the term "input" refers to information that a user inputs into a computer. This may include specified reaction conditions, target polynucleotide sequences that can be divided into polynucleotide fragments, overhangs to be excluded, overhangs to be included, and the number of fragments or overhangs desired. The input parameters are received by the computer.
[0045] As used herein, the term "output" refers to instructions from a computer that allows a user to create a desired target polynucleotide. This may include a set of overhangs with favorable ligation fidelity scores for a particular number of junctions, and / or a complete polynucleotide fragment sequence based on the input of the target polynucleotide. If a polynucleotide fragment sequence is input by the user, the computer output may include pairs of overhangs that avoid internal sites, palindromes, and repetitive overhangs and provide a high overall fidelity score for a particular reaction condition, including cycling conditions, incubation times and temperatures, and recommended enzymes for optimizing ligation fidelity. The computer output may further provide a graphical display of the fragment assembly design and fragment sequences, or provide a link thereto. The computer output may also provide a matrix of ligation frequencies for all combinations of selected overhangs to graphically display the predicted fidelity for the set of selected overhangs and their complements, and a link thereto.
[0046] By default, the tool can provide ligation data with a graphical output showing the general efficiency of each ligation. For example, in Figures 11A-11B, a checkbox can be toggled to display normalized ligation counts. For multiplex ligation assays, the relative ligation frequency was experimentally determined for all 256 4-nt overhangs in a single experiment. The total ligation events for each experiment were normalized to 100,000. In this case, the typical frequency of any single Watson-Crick pair was 300-400 observations per 100,000 ligation events. Further details are described in Potapov et al., Nucleic Acid Research, 46, e79 (2018); Potapov et al., Cold Spring Harbor Laboratory, bioRxiv, doi:https: / / doi.org / 10.1101 / 322297 (2018); and Potapov et al., ACS Synthetic Biology 711, 2665-2675 (2018).
[0047] As used herein, the term "experimental conditions" refers to the selection of ligase, endonuclease and / or other enzymes desired for a workflow and their unit ratios. Conditions also refer to buffers and cofactors in the buffers. For example, the unit ratio of ligase to restriction endonuclease may range from 1:10 to 1:1000, regardless of the type of DNA ligase or type IIS restriction endonuclease selected. Experimental conditions may include salt concentration, temperature and time used to complete ligation of overhangs and may further include cycling conditions for the ligation reaction. Experimental conditions may be selected to reduce assembly time for multiple fragments, improve fidelity scores for a selected set of overhangs, improve activity of the cutting endonuclease while retaining ligase activity and performance, and / or reduce background of misassembly. Experimental conditions may affect removal of mismatches in the target polynucleotide. Although Watson / Crick perfect matches may be preferred, in some cases, single base mismatches within the overhang may provide higher fidelity scores for ordered assembly than perfect matches of bases that do not readily hybridize as predicted from ligation frequency tables. Alternative splicing may also occur during assembly, resulting in mismatches at the junction. Mismatches may be removed using EndoMS or T7 Endo I, or other repair enzymes that identify mismatches and cleave the DNA at the mismatch. The term "experimental conditions" includes ligation conditions, and whether these terms are interchangeable depends on the context.
[0048] As used herein, "ligation frequency" refers to the number of times an overhang is ligated to another overhang out of the total number of ligations (e.g., 100,000 ligations).
[0049] As used herein, the term "ligation fidelity" refers to a numerical assessment of the discrimination (preferential ligation of certain sequences over others) of substrate ligation, including mismatched base pair bias. Ligation fidelity also refers to the ratio of correct ligation events (Watson-Crick ligation products) to incorrect ligation events (mismatch products). In a four-base overhang, there may be no base mismatch (Watson-Crick ligation products), one base mismatch, two base mismatch, three base mismatch, or all four bases mismatch.
[0050] As used herein, the term "ligation fidelity by overhang" or "ligation fidelity score for an individual overhang" refers to the frequency with which an individual overhang and its complement independently ligates to a perfectly complementary overhang compared to all overhangs and their complements in a set. The fidelity score can be calculated by referencing a ligation frequency table that contains individual experimentally defined measurements of the number of ligation events of each overhang to all overhangs of the same length (including itself). The ligation fidelity score of an individual overhang is calculated as the number of ligation events that occur between an individual overhang and its complement compared to the total number of ligation events that occur between (i) the individual overhang and all overhangs and their complements in the set, and (ii) the complements to the individual overhang and all overhangs and their complements in the set.
[0051] As used herein, the terms "ligation fidelity of the entire set" and "overall fidelity score" refer to the predicted ratio of correctly ligated assemblies to incorrectly ligated assemblies based on the individual ligation fidelity scores of each member of a given set of overhangs. The overall fidelity score of a set of overhangs can be calculated by multiplying together the individual ligation fidelity scores of the overhangs in the set.
[0052] As used herein, the term "overall assembly fidelity" refers to the actual number of correctly assembled target nucleic acids compared to the expected number of correctly assembled target nucleic acids. For example, the assembly efficiency of 10 polynucleotides with overhangs can be determined by the number of times all 10 junctions are correctly ordered in a population of target polynucleotides. The assembly fidelity can be greater than 20%, 30%, 40%, 50%, 60%, 60%, 70%, 80%, or 90%.
[0053] The term "ligation efficiency" refers to the number of correct assemblies as a function of time. As used herein, the term "assembly efficiency" refers to the rate at which full-length ligation products (complete target nucleic acids as determined by size or colony formation or sequencing) accumulate in a particular assembly reaction after a particular period of time. Any time unit can be selected that provides an overall average / unit time for the ordered assembly of target polynucleotides. However, ligation efficiency may not be linear over the selected incubation period.
[0054] The term "ligation yield" refers to the number of correct assemblies.
[0055] The term "ligation accuracy" refers to the number of correct end ligations of fragments relative to the total number of assemblies. This can be determined by sequencing.
[0056] The use of the term "ligation" above refers to the product of an assembly that requires a DNA ligase to join the fragments. The use of the term "ligation" below may be due to certain characteristics of the bias and / or fidelity of the ligation event for different ligases that have been found to be variable in ways that may be beneficial or detrimental to the projected assembly.
[0057] The term "ligase" refers to an enzyme that can covalently link two polynucleotides. Many different ligases have been described in the art and are widely known (see Ellenberger et al., Annual Review in Biochemistry, 77, 313-338 (2008); Bauer et al., PLOS ONE, 10, 12: e0145046 (2017)). Ligases for use in the assembly reaction can include ATP ligases and NAD+ ligases such as T4 DNA ligase, T7 DNA ligase, Taq DNA ligase, viral ligases such as Chlorella virus DNA ligase (e.g., PBVC-1 ligase), bacterial ligases such as bacterial LigA (e.g., E. coli DNA ligase) and LigD, archeal ligases such as Thermus thermophilus (Tth) ligase, and eukaryotic ligases such as mammalian Lig1 and hLig3.
[0058] The term "multi-fragment assembly" refers to a plurality of DNA fragments or a set of DNA fragments of any size greater than about 15 nucleotides, either chemically synthesized or within a plasmid in a bacterial library containing plasmids with different inserts. The fragments may all be of similar or identical size, or may have a range of sizes.
[0059] The term "PaqCI" refers to a seven-base cutter restriction endonuclease derived from Paucibacter aquaticus. Endonucleases identified herein as PaqCI include any variant having at least 80%, 85%, 90%, or 95% sequence identity to SEQ ID NO:1.
[0060] In one embodiment, the ordered assembly of multiple polynucleotide fragments onto a single DNA relies on the use of two different enzymes: a Type IIS restriction endonuclease and a ligase. Type IIS restriction endonucleases recognize 4, 5, 6, or 7 bases within the DNA and cleave outside the recognition sequence to provide polynucleotide fragments with overhangs that can be 2, 3, 4, or 5 bases in length. These fragments are ligated when the complementary overhangs anneal, and a ligase seals the linkage.
[0061] A growing number of different type IIS restriction endonucleases have been described that recognize up to six bases and cleave DNA outside the recognition sequence to generate 2-4 base overhangs. Many of the current endonucleases can be found in commercial literature provided by New England Biolabs (NEB), Ipswich, MA (including www.neb.com), including their recognition sequences and the length of the overhangs generated by cleavage. New endonucleases are listed in a database that is regularly updated (see REBASE® from New England Biolabs at www.neb.com). For example, these include: AcuI, AlwI, BaeI, BbsI, BbnI, BccI, BcgI, BciVI, BcoDI, BspMI, BfuAI, BmrI, BpmI, BpuEI, BsaI, BsaXI, BseRI, BsgI, BsmAI, BsmBI, BsmFI, BspCNI, BspMI, BspQI, BsrDI, BsrI, BtgZI, BtsCI, BtsIMutl, CspCI, EarI, EciI, Esp3I, FauI, FokI, HgaI, HpHI, HpyAv, MboII, MlyI, MmeI, MnII, NmeAIII, PleI, SapI, and SfaNI. Only SapI has a 7-base recognition sequence and cleaves DNA to generate 3-base overhangs. All of these endonucleases are available from New England Biolabs, Ipswich, Mass. The recognition sequences and overhangs are described in the NEB, along with reaction buffers, reaction temperatures and storage conditions. Isoschizomer information is also provided.
[0062] Described herein is a novel type IIS restriction endonuclease (PaqCI) that has a 7-base specific recognition sequence, completely cleaves double-stranded DNA, preferably with the aid of synthetic oligonucleotides, and generates 4-base overhangs without star activity (see Figures 1C and 1D). A significant advantage of PaqCI for polynucleotide assembly reactions is that the likelihood of 7-base recognition sequences internal to the polynucleotide fragment sequence is reduced. Internal recognition sequences are undesirable because they would cleave the polynucleotide fragments, thereby disrupting the ordered assembly of the fragment set.
[0063] The recognition sequence for PaqCI is added to the ends of the fragments by primer-dependent amplification or chemical nucleic acid synthesis. The sequences flanking the restriction endonuclease recognition sequences form overhangs. Preferably, these are selected to optimize ligation of an ordered set of fragments comprising a target polynucleotide or large DNA.
[0064] Although the use of PaqCI is described herein in the context of ordered polynucleotide fragment assembly, the enzyme may also be used in a variety of other bioengineering methods and analyses of genomic DNA, including chromatin, where endonucleases with recognition sequences of six or more bases are preferred.
[0065] In one embodiment of the present invention, the ordered assembly of multiple polynucleotide fragments into a single DNA may rely on the use of ligase, but no restriction endonuclease is required. If the polynucleotide fragments with designed overhangs are generated by chemical synthesis instead of endonuclease cleavage, no restriction endonuclease is required. However, the advantage of the two enzyme system is that the polynucleotide fragments can be inserted into a plasmid that is maintained in bacteria and can be stored indefinitely for future use. Optionally, the bacteria can be easily grown to generate the desired amount of substrate for endonuclease cleavage and ligation.
[0066] Ordered assembly of multiple polynucleotide fragments that relies on ligation of annealed overhangs has been significantly improved by systematic analysis of overhang ligation frequency, bias, and fidelity. Factors identified include overhang length, number of different overhangs in a set of overhangs, GC content of the overhangs, bases occurring at the ends of the overhang sequences, ligation reaction conditions, and type of restriction endonuclease generating the overhangs (see WO2020 / 081768).
[0067] In the present embodiment, sequence preferences of various ligases with robust end-joining activity were surprisingly identified and found to contribute significantly to the frequency and fidelity of ligation products. Differences and similarities between various ligases were identified, including the degree of mismatch tolerated between annealed overhangs and preferences for specific patterns of A, T, G, and C bases in the overhangs.
[0068] The ligases described in the examples are all end-linking ligases that catalyze the formation of a phosphodiester bond between the 3'-hydroxyl of one DNA strand and the 5'-phosphorylated end of another DNA strand. They all contain at least two domains corresponding to a nucleotidyl transferase domain (NTase) with a catalytic lysine residue and an oligonucleotide binding domain (OBD) with a DNA binding surface. The ligases also optionally contain a third domain. For example, both T4 DNA ligase and T3 DNA ligase contain an N-terminal DNA binding domain, while human ligase 3 (hLig3) contains an N-terminal poly ADP-ribose polymerase-like zinc finger domain and Chlorella virus DNA ligase (PBCV-1) contains a latch domain. However, T7 DNA ligase does not contain a third domain. Without wishing to be limited by theory, the presence or absence of a third domain may play a role in the bias, promiscuity and / or fidelity of ligation.
[0069] Each of the 256 combinations and permutations of 4-base overhangs was analyzed for end-ligation activity. A 256×256 sequence matrix was constructed from sequencing data obtained from ligated overhangs using the assays outlined in FIG. 2A and FIG. 2B. (See FIG. 3Ai / 3Aii to 3Hi / 3Hii.) Data from these assays were fed into a computational design tool described in FIG. 6, FIG. 7, FIG. 8, and FIG. 9A and FIG. 9B that allows the user to select a set of optimized overhang sequences for the ordered assembly of a set of polynucleotide fragments.
[0070] The computer tool described in WO2020 / 081768 provides access to a set of overhangs optimized based on their annealing patterns. Restriction endonucleases can be selected from a menu of options for cleavage to generate the overhangs. Ligation frequency and fidelity metrics are provided for a variety of overhangs using a single ligase, namely T4 DNA ligase, under assembly conditions that can be selected from drop-down menus, including buffer options and temperature and incubation time options.
[0071] The ligation data described herein and in WO2020 / 081768 captures the preferences of ligase substrates, further enhancing the accuracy of the aforementioned assembly options. This is particularly important when a large number of polynucleotide fragments (more than about 20 fragments) are used for the ordered assembly of large DNA. The computational tool provides a suitable user interface to inform the user about the predicted efficiency (frequency) and fidelity profile for any fragment overhang or set of fragment overhangs under various experimental conditions. The data obtained on the ligation preferences of different ligases expands the menu of experimental conditions. The interfaces for the ligation fidelity viewer, GetSet, and SplitSet, including drop-down menus, allow the user to select a suitable ligase for in silico designing fragments with overhangs from large DNA sequences. Alternatively, the user can select the appropriate ligase for a given set of overhangs. Different sequence preferences for ligation resulting in ligation frequency and mismatch frequency as well as different fidelity profiles add an additional layer of refinement and efficiency for multi-fragment assembly. Modifications to standard ligase buffers that affect ligase activity, such as polyethylene glycol, are also described herein.
[0072] Tools and methods are provided to allow assembly of a greater number of pieces with higher fidelity in the assembled sequence and a higher frequency of bacterial colonies transformed with destination vectors containing assembled DNA or packaged virus that infect bacterial lawns. Benefiting from the improvements, an intact T7 viral genome has been assembled from 50 pieces. The newly synthesized virus has been shown to generate viral plaques in bacterial lawns. The availability of ligase data provides an improvement of at least 10%, 20%, 30%, 40%, or 50% more colonies in 24-piece and 50-piece assemblies than would otherwise occur.
[0073] The ability to assemble a small number of polynucleotide fragments (e.g., less than 10 fragments) into a larger DNA can be done relatively efficiently without additional refinement. However, there are advantages associated with assembling a larger number of smaller fragments (e.g., more than 10 fragments, or up to 20 fragments, or up to 50 fragments, or up to 100 fragments, or more, etc.) to generate DNA of similar size. Such advantages include a lower error rate occurring with small synthetic oligonucleotides than with large synthetic oligonucleotides, and the ease of stably maintaining bacterial clones with small size plasmid inserts, which can be stored and used as needed for various assembly projects to generate larger DNA. This avoids toxicity issues that may occur when larger fragments are expressed in bacterial clones, affecting the viability of the clones and thereby reducing the quality of the library of stored fragments. Furthermore, assembly reactions involving a large number of fragments provide the ability to generate variant polynucleotide libraries that may be useful for vaccine development, car-T therapy, and phage-based antibiotic development, as outlined in Figures 15 and 16 and in the Examples.
[0074] A novel type IIS restriction endonuclease - PaqCI PaqCI is characterized as a protein having at least 80% sequence identity to SEQ ID NO: 1. As used herein, PaqCI is intended to encompass variants having at least 90%, at least 92%, at least 95%, at least 99% sequence identity to SEQ ID NO: 1. PaqCI relies on multiple subunits that interact with two recognition sites to cleave a single target site on each strand of a DNA duplex.
[0075] PaqCI sequence, 510 aa (SEQ ID NO:1):
[0076] [ka]
[0077] The recognition sequence of PaqCI is (5'-CACCTGC-3' / 3'-GCAGGTG-5'), which asymmetrically cleaves 4 bases from the recognition sequence in the 3' direction and 8 bases from the complement of the recognition sequence in the 5' direction, resulting in a 4-base overhang (see Figure 1C).
[0078] Also described herein are PaqCI activator oligonucleotides that have been found to improve the activity of PaqCI. Activator oligonucleotides are synthetic self-complementary single-stranded oligonucleotides that are folded to include a double-stranded DNA region and a single-stranded DNA loop, e.g., a hairpin structure. The advantages of a hairpin over two single strands include more complete annealing, since the two ends of a single synthesized DNA strand are at exactly the same concentration.
[0079] The double-stranded region of the activator oligonucleotide contains a binding (recognition) sequence for PaqCI, and the oligonucleotide contains 3' and 5' ends that cannot be ligated and cannot be cleaved by PaqCI, meaning that the double-stranded portion of the oligonucleotide does not extend far enough beyond the recognition site to provide a cleavage site for PaqCI. The self-complementary oligonucleotide containing the double-stranded region and the loop is preferably less than 100 nucleotides in length, contains a recognition sequence for PaqCI (5'CACCTGC / 3'GTGGACCG) and extends 0-4 bases or 1-4 bases or less downstream from the 5' recognition sequence. One non-natural extension of the blocking moiety on each strand can be added so that there is no properly positioned phosphodiester bond in the double-stranded region for the enzyme to cleave. Alternatively, the activator oligonucleotide may contain a non-cleavable linkage. The 5' and 3' ends of the oligonucleotide may be flat or may be interrupted by 1, 2, 3, 4, 5, 6 or more nucleotides, and either the 3' or 5' end may be interrupted. The length of the oligonucleotide loop is not critical and may in some cases be 4 to 20 nucleotides. The double-stranded region may be 10 to 50 base pairs in length, such as 10 to 30 base pairs in length, for example 15 to 30 bases.
[0080] The activator oligonucleotide has a non-ligatable 3' or 5' end that cannot be ligated to another substrate (polynucleotide fragment or activator oligonucleotide) by T4 DNA ligase or other ligase in T4 DNA ligation buffer or other suitable ligase buffer. Examples of non-ligatable 3' and 5' ends are a 3' end that does not contain a 3' hydroxyl and a 5' end that does not contain a 5' phosphate; a 3' end that contains a 3' phosphate and a 5' end that contains a C3 spacer; or alternatively, a ligation block at the 3' end such as 3' dideoxy-C, 3' C3 spacer (C3-OH), C6 spacer or 3' amino linker (C6-NH2) and a ligation-inhibiting modified base at the 5' end such as inverted dideoxythymine (invddT). Thus, ligation of the activators to each other or to the polynucleotide fragment is prevented.
[0081] The reaction mixture containing PaqCI also contains one or more activator oligonucleotides for adding to the target double-stranded DNA intended for cleavage.
[0082] It is within the scope of these embodiments to use a mixture of PaqCI and one or more variants of PaqCI in the same reaction mixture that includes one or more activator oligonucleotides. In certain embodiments, methods are provided that further include one or more ligases. In certain embodiments, PaqCI can be used in a mixture with other restriction endonucleases having different or the same specificity.
[0083] The amounts of PaqCI and activator were optimized to be within a range that resulted in substantially complete cleavage of the DNA substrate by PaqCI but did not result in star activity. The ratio of PaqCI to activator was found to be more significant for optimizing enzyme activity than the ratio of activator to recognition sites on the target oligonucleotide. Insufficient concentrations of activator relative to PaqCI resulted in incomplete cleavage of the target DNA and star activity. Too much activator resulted in incomplete cleavage. Without wishing to be limited by theory, it is believed that the incomplete cleavage was the result of PaqCI binding exclusively to the activator molecule instead of the target DNA.
[0084] The optimal amount of activator for a given amount of PaqCI may vary depending on its intended use. A standard restriction digest with PaqCI without a complex assembly reaction in the same tube can be accomplished using 1 μl of enzyme (10 U) and 1 μl of activator (20 pmoles). In these reactions, once the DNA substrate is cleaved, it does not readily reassemble.
[0085] However, when PaqCI is used in a multi-fragment assembly method, the overhangs generated by endonuclease cleavage may be able to reanneal and ligate to reconstitute the original recognition site. In this dynamic situation, any one DNA cleavage site may need to be cleaved more than once throughout the assembly reaction. As a result, the ratio of endonuclease to activator was altered according to the number of fragments in the assembly reaction.
[0086] As a result, it was determined that 0.75 pmoles to 9 pmoles of activator per unit of PaqCI endonuclease (15 nM to 180 nM in a standard 50 μl reaction volume) was preferred, with some small amount of star activity being observed at less than 0.75 pmoles of activator per unit of PaqCI, but with the onset of inhibitory activity being observed at 10 pmoles of activator per unit of PaqCI. In certain embodiments, the range may be selected from any of 0.75 picomoles to 9 picomoles of activator per unit of PaqCI, 1 picomoles to 7.5 picomoles of activator per unit of PaqCI, 1 picomoles to 5 picomoles of activator per unit of PaqCI, 1.5 picomoles to 7.5 picomoles of activator per unit of PaqCI, 1.5 picomoles to 5 picomoles of activator per unit of PaqCI, 1.5 picomoles to 4 picomoles of activator per unit of PaqCI, 2 picomoles to 5 picomoles of activator per unit of PaqCI, or 2 picomoles to 4 picomoles of activator per unit of PaqCI.
[0087] One unit is defined for this ratio as the amount of enzyme required to digest 1 μg of λ DNA in 1× rCutSmart™ buffer (50 mM potassium acetate, 20 mM Tris acetate, 10 mM magnesium acetate, 100 μg / ml recombinant albumin (pH 7.9 at 25° C.) in a total reaction volume of 50 μl for 1 hour at 37° C. The unit definition and buffer description are not intended to be limiting, but rather serve as a guide to develop an appropriate activator / PaqCI ratio. Other buffers can be used depending on the application, including the selected ligase. For example, for DNA fragment assembly methods, commercially available T4 DNA ligase buffer may be preferred. In one embodiment of the assembly method, the standard reaction volume is 50 ul and contains 1-8 pmoles of activator (20 nM-160 nM) per unit of PaqCI endonuclease or 10-80 pmoles of activator (200 nM-1600 nM) per 10 units of enzyme. The DNA in the reaction mix was 1 ug of lambda DNA. In another embodiment, the activator concentration is 20 uM and the enzyme concentration is 10 units / ul, such that the optimal enzyme and activator in a 50 ul reaction yields a 1:1 ratio using 1 ul each. For typical PaqCI reaction conditions, this results in complete cleavage of the DNA substrate recognition site, even though the concentration of the substrate site varies depending on the DNA being cleaved. For multi-fragment ligation for polynucleotide assembly, where the concentration of substrate sites is high compared to typical reaction conditions, a lower ratio of activator to enzyme (e.g., 1:2, 1:3, 1:4, or 1:5) can be used. By lowering the amount of activator, the optimal ratio of enzyme to total recognition sites (substrate plus activator sites) can be maintained.
[0088] The recommended component concentration ranges described herein were also determined for multi-fragment assembly methods containing different numbers of insert clones resulting in simple to complex assemblies.
[0089] For example, PaqCI or a variant thereof may be combined with an activator, a ligase, and multiple DNA substrates in a reaction mix. In one embodiment, the DNA substrates are contained in a plasmid containing a PaqCI recognition sequence at the insertion site with adjacent plasmid sequences designed for ligation assembly of the substrates. During multi-fragment assembly, every insert and every destination plasmid has an assembly-active DNA fragment flanked by two sites. The reaction mix may be incubated for a time and temperature suitable for endonuclease cleavage and ligation of the fragments (e.g., 37°C and 60°C for 30-60 ligation cycles, where each cycle is 1-5 minutes depending on the number of fragments in the mix). The desired reaction product is a large DNA molecule formed from multiple DNA substrates. Fragment assemblies of different levels of complexity require different levels of the PaqCI and DNA ligase described above. As assembly reactions increase in complexity, more units of enzyme are required for maximum performance; for example, using T4 DNA ligase, 2.5-20 U of PaqCI can be used along with 200-800 U of ligase, and for assembly of 20 or more fragments, an upper limit of 10-20 U of PaqCI and 400-800 U of DNA ligase is preferred.
[0090] PaqCI cleaves to completion and does not have a star activator when combined with an activator (see FIG. 1D). It has greatly improved performance when compared to AarI (see FIGS. 1A and 1B).
[0091] In embodiments, a kit is provided that contains a mixture or reagents in one or more containers, the reagents including PaqCI or a variant thereof ("PaqCI") and an activator molecule. The kit may further include a ligase. The kit may include the reagents in a reaction buffer, or one or more of the reagents may be lyophilized and / or immobilized together or separately on a suitable substrate, such as a bead or polymer matrix. The kit may further contain a reaction buffer in a separate container for addition to the reagents.
[0092] Ligation frequency and fidelity for different ligases Multi-fragment assembly can be achieved by combining PaqCI with a selected ligase to generate fragments with 4-base overhangs. Ordered assembly depends on the fidelity of overhang annealing and the promiscuity of ligation by the ligase of all annealed overhangs, which in turn depends on the conditions of ligation, including the number of fragments that form intact contiguous DNA. An embodiment of the present invention establishes the role of different ligases in unique ligase preferences for ligation-associated annealing.
[0093] T4 DNA ligase is a standard ligase for end ligation and large DNA assemblies. However, it was unclear whether this ligase has sequence preferences that contribute to the variable ligation profiles observed for end-joined fragments with certain 4-base overhangs. Furthermore, it was unclear how T4 DNA ligase compared to other ligases in this regard.
[0094] A method was developed herein to analyze this question and use the results of the analysis to improve overhang selection, minimize bias, and optionally enhance the yield and optimize fidelity of assembled fragments. A detailed analysis of the properties of five ligases is provided in the figures and examples (T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, hLig3 DNA ligase, and PBCV-1 DNA ligase). Furthermore, certain improvements observed by adding PEG and / or aprataxin to the ligation buffer are described in the examples.
[0095] In Example 1, the ligation frequency, bias and fidelity profiles of DNA ligases T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, PBCV-1 DNA ligase, and hLig3 were determined using libraries of end-joined hairpin DNA substrates containing degenerate 5'-4 base overhang ends. The ligation products of these libraries were analyzed by sequencing. The number of reads for each overhang provided a value for ligation efficiency; sequence bias for each ligase was inferred from the relative frequency of each overhang appearing across all ligation products.
[0096] The hairpin substrates in the assay present a complex equilibrium system that mimics the actual assembly of multiple fragments, where ligation requires a ligase to find complementary ends of the fragments. Rapid conversion to ligated products is predicted when only two Watson-Crick binding partners are present in the reaction. The assay method provided a depth of information not available by examining each overhang separately, allowing for a more rapid assessment of fidelity and bias than would have been possible by testing each pair simultaneously. Raw data on the frequency of ligation of each of all complementary 4-nt overhangs are presented in a heat map (matrix) in Figures 3Ai / Figure 3Aii-3Hi / Figure 3Hii for different ligases under the same assay conditions. The data showed that the ligation frequency and fidelity of the libraries varied significantly between the ligases tested. T4 DNA ligase, T3 DNA ligase, and hLig3 all resulted in ligation products higher than 55%. PBCV-1 ligase had a slightly lower yield (50%). T7 DNA ligase had by far the lowest yield, reaching only 20% ligation product.
[0097] A variation in the degree of overall bias and disparate preferred sequences between the ligases was observed (see Figures 3Ai / 3Aii-3Hi / 3Hii). T7 DNA ligase showed the highest degree of sequence bias in favor of ligating perfect matches of bases in 4-nt overhangs. All other ligases tested had much tighter distributions of ligation frequencies, but there were differences in how tightly the data points clustered around the mean (see Figure 4A). Both T4 DNA ligase and hLig3 showed the least amount of bias, with a range of values less than 2-fold smaller than T7 DNA ligase. PBCV-1 and T3 had similar average ligation frequencies, but a slightly larger range of ligation frequencies was observed.
[0098] The ligases tested here showed extreme fidelity with T7 DNA ligase exhibiting the highest fidelity (89% correct ligation products), whereas hLig3 had the lowest fidelity (56% correct ligation products). T4 DNA ligase exhibited moderate fidelity (72% correct ligation products). T4 DNA ligase, T3 DNA ligase, PBCV-1 ligase, and hLig3 had a wide range of fidelity with respect to individual overhang sequences, with some overhangs having very few mismatch ligation events and others having frequent mismatch ligation (Figure 4B). For many overhangs, even when presented with every possible partner, the ligation products were almost exclusively with the Watson-Crick partner. When specific mismatch base pairs tolerated by each ligase occurred, these were more common at the 5'-terminal nucleotides ("ends") than at the center of the overhang ("middle"). Although some frequent mismatches, notably the G:T mismatch, were common among all tested ligases, there were also different mismatch pairs observed between the ligation products of each ligase.
[0099] For example, when T4 DNA ligase was presented with all possible ligation partners, some overhangs paired with their Watson-Crick partners in more than 90% of the ligation products (e.g., AAAA, AAGA, ACAA, GAAA). Other overhangs ligated to partners containing at least one mismatch; some overhangs paired with mismatch-containing partners more than 60% of the time (e.g., GGCG, GGCC, GGGC, GGGG).
[0100] For certain multi-fragment assemblies, it is desirable to maximize Watson-Crick matches to provide a completely accurate sequence, but in other situations it may be desirable to introduce errors during fragment assembly to establish highly variable DNA. For example, hlig3 may be selected to increase the chance that a fragment in a set of fragments will ligate to another fragment in the wrong order due to ligase promiscuity. Alternatively, if maximum fidelity of assembly is desired and frequency of ligation is not critical, T7 DNA ligase may be the enzyme of choice. However, as shown below, additives to the ligation buffer, such as PEG, can enhance the frequency of ligation somewhat without significant loss of fidelity, such that T7 DNA ligase may be the ligase of choice for 20+ fragment assembly workflows where this ligase would otherwise be less desirable.
[0101] Addition of the ligation enhancer PEG not only reduces ligation bias but also fidelity Polyethylene glycol (PEG) (Millipore Sigma, Burlington, MA) can also be used to enhance ligation. The examples show that PEG with molecular weights ranging from 600 to 6000 enhances DNA assembly. For example, PEG MW can be selected from 500, 600, 800, 1000, 1500, 2000, 2500, 3000, 3500, 4000, 4500, 5000, 5500, 6000, 6500, 7000, 7500, 8000, 8500, 9000, and 10,000. Example 4 shows results for PEG600, PEG3350, and PEG6000 at 6.8% w / v.
[0102] The addition of PEG increased the overall library yield for both T4 and T7 DNA ligase (61%-73% and 20%-45%, respectively), with a slight decrease in the yield for hLig3 (77%-72%). The addition of PEG moderately decreased the overall fidelity of the multiplex ligation reaction for T4 DNA ligase from 72% correct ligation events in the absence of PEG to 67% in the presence of PEG (see Table 2, Figures 5A-5E). The addition of PEG reduced the fidelity by the same amount regardless of GC content, except for overhangs with 100% GC content, which saw no change in average fidelity. The fidelity of T7 DNA ligase was also significantly decreased in the presence of PEG, from 89%-78%, and the overall ligation fidelity of hLig3 was slightly decreased upon the addition of PEG (56% and 51%, respectively) (Table 2). Notably, the addition of PEG did not change the identity of the specific mismatch tolerated for any of the ligases tested. The addition of PEG also resulted in a moderate change in the bias of T4 DNA ligase. A boost in efficiency was observed for ligation of overhangs with less than 50% GC content. Interestingly, the extreme bias of T7 DNA ligase was reduced by the addition of PEG, and overhangs that had not previously ligated at all had significant product accumulation (Figures 5A-5E). Results obtained with PEGs of different sizes are provided in Table 6.
[0103] Typical reaction conditions and additives may provide insight into modifications that may affect different DNA ligases and improve the outcome of certain applications. For applications such as cloning or adapter ligation, the boost in ligation product yield from the addition of PEG is likely to outweigh the moderate loss of fidelity for T4 and T7 DNA ligases. However, for applications involving highly complex multi-fragment assemblies, the loss of fidelity observed when adding PEG may require more consideration of the specific overhangs used to limit potential mismatch ligation between specific sets of overhangs. The addition of PEG makes T7 DNA ligase a more attractive candidate for large multi-fragment assemblies. The observed increase in efficiency with additional overhangs expands the pool of efficient potential overhang sequences, but the small loss of fidelity is tolerated due to the high overall fidelity of this enzyme.
[0104] Data-optimized assembly design The Ligase Fidelity Viewer, GetSet and SplitSet tools described herein and in WO2020 / 0181768 provide data-optimized assembly designs that greatly improve the success of ordered assembly of fragments. These tools rely on menu selections to assist the user. Each of the experimental conditions described below is derived from a 256x256 data matrix of ligation frequency and fidelity. A computational tool can use this data to generate overhangs that are optimized for the desired number of fragments and overhang type.
[0105] Selection of a suitable set of overhangs for assembling a selected number of polynucleotide fragments into a target polynucleotide can be accomplished by these tools to test and arrange in a suitable order useful overhangs that cannot be achieved by mental acts. In one embodiment, identification of suitable overhangs includes one or more steps, such as: not including palindromic overhangs, not including duplicated overhangs, not including overhangs with 3 consecutive bases, e.g. ACC A and ACC Contains G; does not contain more than two bases at the same position, e.g., ( A C G C and A T G Certain rules were applied that included 0% GC overhangs and 100% GC overhangs (Nilsson et al., Nucleic Acids Res., 10, 1425-1437 (1982); Goffin et al., Nucleic Acids Res., 15, 8755-8771 (1987); and Wu et al., Gene, 76, 245-254 (1989)). However, the optimized data set determined by the data-optimized assembly design exemplified herein does not have such rule-based restrictions and provides more flexibility in the selection of overhangs. For example, temperature plays a role in the overall degree of overhang annealing due to AT and GC content. The preferred positions of A, T, G or C bases in the overhang were found to vary depending on the ligase used. Although very few mismatches exist within the 4-base overhang, different preferences for various ligases are observed for the bases at either end of the overhang.
[0106] The embodiments describe how to obtain optimized datasets by profiling end-joining hybridization and ligation fidelity and bias to predict a highly accurate set of connections for ligation-based DNA assembly methods. This represents a significant improvement over prior art rules that restrict users to a limited number of 4-base overhangs, which is particularly restrictive when sequences cannot be chosen arbitrarily (e.g., when assembly must cut within a coding sequence). Application of ligation fidelity profiles allowed for informed selection of junctions, enabling highly flexible assembly designs using more than 20 fragments in a single reaction.
[0107] In one embodiment, the computer design tool provides for the selection of overhangs after a user inputs various requirements. For example, the computer tool receives a selection or input of experimental conditions under which the assembly reaction is expected to occur (see, e.g., Table 1). The experimental conditions drive the selection of ligation frequency tables and the determination of ligation fidelity of individual overhangs, which are accessed by the computer.
[0108] The computer tool also receives a number, e.g., (i) a desired number of overhangs for the assembly reaction, (ii) the length of the overhangs, (iii) a nucleotide sequence for assembly, (iv) a set of intervals at which the nucleotide sequence of (iii) can be cleaved, and after running the method, receives as output the set of identified overhangs and / or receives a set of fragments for assembly where the ends of the fragments are defined by the set of identified overhangs depending on the information entered into the interface.
[0109] The method may further include receiving instructions regarding ligation conditions for ligating the overhang or the set of fragments containing it. The ligation conditions may include one or more of the desired ligase, buffer conditions, such as salt concentration, temperature, temperature range and / or thermal cycling time and temperature (which can be used for cleavage and ligation), and restriction endonucleases, if used to generate the overhang. When the ligation conditions (in addition to the desired number of overhangs for the assembly reaction and the overhang length) are received at the interface, an overall fidelity score is calculated for the overhangs in the ligation performed at the specific ligation conditions. Thus, a ligation frequency table corresponding to a specific condition may be selected from a plurality of ligation frequency tables, each for a different ligation condition. After selecting a ligation frequency table, a ligation fidelity score may be calculated using the data in the table.
[0110] In some embodiments, the number of overhangs may be in the range of 5 to 200, e.g., 10 to 100, e.g., 10 to 50, although the number of overhangs may be outside these ranges in some circumstances. The length of the overhang may be 2, 3, 4 or 5 nucleotides, with the length of the overhang being limited only by the choice of the restriction endonuclease or other means for generating the overhang and the frequency and fidelity of possible ligation reactions. The choice regarding the preferred length of the overhang may be influenced by the number of overhangs possible for any combination of nucleotides in the overhang, which should exceed the number of fragments to be joined. After receiving the desired number of overhangs for the assembly reaction and the length of the overhang, the computer provides a set of overhangs from the overhang table of the selected length (e.g., 2, 3, 4 or 5 bases). For example, if the user inputs into the computer 20 overhangs, each 4 bases long, the computer will output a set of 20 unique overhangs that did not contain any duplications, complements, palindromes (e.g., GATC), or excluded sequences. GATC is an example of a palindromic sequence because its reverse complement is GATC. Palindromes should be avoided because any one fragment with a palindromic end may anneal to another identical molecule, resulting in disruption of the ordered assembly. The interface may allow receiving a list of one or more overhangs to be excluded or included. Overhangs to be excluded may be selected due to low fidelity or frequency in the ligation profile, or because the overhang was used elsewhere in the reaction. Overhangs to be included may be selected based on experimental findings of their high fidelity and / or frequency values.
[0111] As will be appreciated, the overhang ligation frequency table is stored in a computer memory and may include all possible overhangs of a desired length or a subset thereof. For example, for a 4-base overhang, the overhang table may contain 256 4-base sequences, and for a 3-base overhang, the overhang table may contain 64 3-base sequences. The overhangs may be selected in any particular order. For example, in some embodiments, the overhangs may be selected randomly, while in other embodiments, the overhangs may be selected in a prescribed order.
[0112] Based on the user's selection from the menu, the computer calculates a ligation fidelity score for each individual overhang and its complement in the set. For example, if there are 20 overhangs in the set, there should be 20 ligation fidelity scores, with the ligation fidelity score of each individual overhang representing the frequency with which the individual overhang and its complement independently ligates to a perfectly complementary overhang to all overhangs and their complements in the set. For example, if a particular overhang and its complement ligates to each other with perfect complementarity 90% of the time to all overhangs and their complements in the set, the overhang may have a calculated ligation fidelity score of 0.9. If a particular overhang and its complement ligates 95% of the time, the difference suggests a calculable rate of ligatable mismatches. These values are obtained from the described ligation fidelity / frequency table obtained by the method described herein, with each entry consisting of an individual experimentally defined measurement of the number of ligation events taken into account in different experimental conditions.
[0113] Ligation Conditions Ligation conditions can be selected using a drop down menu where the ligase options laid out in the drop down menu include various ligation frequency tables. An example where the choice of ligase is added to experimental conditions found to affect ligation efficiency, fidelity and yield affects experimentally determined values for overhang ligation frequency and fidelity for ordered assembly of fragments.
[0114] Temperature conditions for ligation, including static temperature and cycling between high and low temperatures using drop-down or touch-down cycling. These terms are explained in Example 2. (a) Number of fragments for ordered assembly, (b) the length of the protrusion; (c) Restriction endonuclease / ligation reaction incubation time; (d) types of restriction endonucleases and ligases; (e) type of buffer, including salt concentration; (f) cofactors such as crowding agents, repair enzymes and / or deadenylases (see also Tables 4 and 5); (g) Selection of a ligase; (h) acceptable ligase bias according to or against a particular tetramer sequence, mismatch tolerance, and fidelity profile; (i) Specific 4-base sequences, mismatch tolerance, and intolerant ligase bias according to or against the fidelity profile.
[0115] In one embodiment, the experimental condition pull-down menus in the Ligation Fidelity Viewer, GetSet, and SplitSet user interfaces in FIGS.
[0116] [Table 1]
[0117] Each entry had a 256x256 data set that was input into the Data Tool for integration into the calculation of optimal overhangs. Example 1 details how the data was collected for the comparative study of the five ligases.
[0118] The ligation frequency table for 4-base overhangs should have experimental values for all possible combinations of overhangs, i.e., 256x256 / 2 data points, with each value indicating the frequency of ligation of two overhangs under defined experimental conditions. Details of how to obtain this data are described in Example 1 and in Potapov, V. et al. (2018), ACS Synth. Biol., vol. 7, pp. 2665-2674; Potapov et al., Nucleic Acid Res 2018, 46 e79; Potapov et al. (2018) BioRxiv; Pryor, JM et al. (2020) PLoS One, e8592; Pryor, JM et al. (2020) BioRxiv, e4019. The ligation fidelity score of an individual overhang can be calculated as the number of ligation events that occur between an individual overhang and its complement compared to the total number of ligation events that occur between (i) the individual overhang and all overhangs in the set and their complements, and (ii) the complements to the individual overhang and all overhangs in the set and their complements.
[0119] An overall fidelity score for the set of overhangs can then be generated based on the ligation fidelity scores calculated for each of the individual overhangs, as output above. In some embodiments, the individual ligation fidelity scores may be multiplied together to obtain an overall fidelity score. For example, if there are 20 overhangs, each with a fidelity of 0.950, then the overall fidelity score for the set of overhangs will be 0.36 (i.e., 0.950). 20). In some embodiments, this calculation may additionally weight the overhangs according to how efficient they are at ligating to their complement. For example, in some cases, two overhangs may have equal fidelity, but under the conditions used, one ligates to its complement more efficiently than the other. In this case, the overhang that ligates more efficiently may have a higher weight than the other overhang. Thus, in some embodiments, an overall fidelity score may be calculated using (i) the ligation fidelity scores calculated for each of the individual overhangs, and (ii) the yield of each of the individual overhangs ligating to a perfectly complementary overhang.
[0120] After calculating the overall fidelity scores of the selected set of lobes, the process may be repeated for another set of lobes to calculate multiple overall fidelity scores for each of the different sets of lobes. In this step, the sets of lobes selected in the iteration step are different from each other (and different from the first set of lobes). Also, in the iteration step, the selection may be random or in a prescribed order. In some embodiments, these steps may be repeated using Monte Carlo simulation. In this method, at least 100, at least 1,000, or at least 10,000 overall fidelity scores may be generated for each different set of lobes. This part of the method is repeated until an overall fidelity score has been assigned to all possible combinations of lobes or until one or more lobes have been identified with an overall fidelity score above a threshold.
[0121] After the overall fidelity score is calculated, the method may include identifying a set of saliencies having an appropriate overall fidelity score (examples of sets of saliencies are provided in Table 7 and Example 5). In some embodiments, the identified set of saliencies may have an overall fidelity score in the top 50%, top 20%, top 10%, or top 5% of the overall fidelity scores. In some embodiments, the identified set of saliencies may have the highest overall fidelity score or a score within the top 10% or top 5% of the highest fidelity scores. The selected set of saliencies may be output from the computer, for example, to a display (see Example 5 and Figures 9B-9D).
[0122] In further embodiments, the method may include a user inputting one or more of the following into an interface: (i) a desired number of overhangs for the assembly reaction, (ii) a length of the overhangs; optionally (iii) a nucleotide sequence for the assembly, (iv) a set of intervals at which the nucleotide sequence of (iii) may be cleaved, running the method and receiving as output the set of identified overhangs and / or receiving a set of fragments for assembly, where the ends of the fragments are defined by the set of identified overhangs depending on which information is input into the interface. The method may further include receiving instructions regarding ligation conditions for ligating the overhangs or the set of fragments containing them, and optionally thermal cycling conditions for generating the fragments and ligating them together. In these embodiments, the method may include generating a set of double-stranded nucleic acids having a set of overhangs with an overall ligation score above a threshold and their complements, and then ligating the fragments together in a single reaction to generate an assembly, where the overhangs determine the order of the fragments in the assembly. Ligation may be performed by overhang-directed ligation, as described in more detail above and / or below. As will be apparent, the method may further include receiving selected experimental conditions for ligation.
[0123] Implementation of the above embodiment is exemplified by the Ligation Fidelity Viewer, GetSet, and SplitSet, which are described in detail in WO2020 / 081768. The user interfaces for each of these applications are shown in Figures 9A-9D. The ligase data provided in the examples is an additional feature of the experimental conditions as described above, which allows for the refinement of the optimized overhang set. This is particularly useful for large overhang sets with corresponding large fragment sets for ordered assembly.
[0124] In GetSet (see FIG. 9C), select the overhang length, enter the total number of overhangs, enter the overhangs required, you may also add omitted overhangs, and select experimental conditions including the use of PaqCI and choice of ligase. GetSet then provides a set of overhangs that are optimal for the specified ligation conditions.
[0125] In SplitSet, the first step may include receiving a nucleotide sequence of the assembly of interest and a set of intervals (e.g., at least 5, at least 10, at least 20, or at least 30, up to 50 intervals) at which the nucleotide sequence can be cut (in addition to the desired number of overhangs and the length of the overhangs for the assembly reaction). The input sequence may be any sequence of at least 500 bases in length, for example, but sequences as short as 25 nucleotides can be selected, provided that a type IIS restriction endonuclease recognition sequence is present at the beginning and end of the interval. For example, the method may include receiving a sequence and multiple sets of start and end coordinates, where each set of start and end coordinates defines an interval at which the sequence can be cut. In these situations, only the overhangs that lie within the interval may be selected from an overhang table, such that each interval is jointly represented by the selected overhang. A non-redundant set of subsequences is then identified in an interval of length equal to the length of the received overhang. These subsequences may be stored as an overhang table itself, or only sequences from a non-redundant set of subsequences may be selected from the overhang table (see Tables 8 and 9). The interval may be input to the computer by the user, for example, by entering the interval into an interface (see FIG. 9D). Alternatively, the user may input the sequence and specify the number of fragments desired. In these embodiments, the algorithm may determine approximate positions at which the input sequence may be split to generate the desired number of fragments, and then specify an interval (which may be, for example, 10-50 or 10-100 nucleotides long) that includes the approximate positions. The interval may be processed as described above. In these embodiments, the method may further include splitting the nucleotide sequence of the assembly at the identified overhangs, thereby generating a set of fragments of the assembly, where the ends of the fragments are defined by the identified overhangs. The SplitSet interface is shown in FIG. 9D. The desired overhang length is provided by selecting an item in the menu.Next, similar to the Ligation Fidelity Viewer, ligation conditions are selected, nucleotide sequences are entered, and fragment numbers are entered, and the computer then provides the result of an optimized set of fragments for ordered assembly.
[0126] The embodiments provided herein are intended to allow a computer user to view, through a graphical representation, the ligation fidelity profile predicted from a given set of fragment overhangs under selected experimental conditions. Each of these features may be modified by adjusting any of the parameters described herein to provide a modified graphical representation, and determining whether the modification improved the ligation fidelity profile of a selected number of overhang sequences using the graphical representation of the deviation from a perfect score obtained for the set of overhang sequences.
[0127] Other embodiments are provided herein to allow a computer user to rapidly and efficiently obtain from a computer an optimized set of overhang sequences suitable for assembling multiple nucleic acid fragments into a target polynucleotide. The optimization may rely on two or more databases of ligation fidelity and ligation efficiency (frequency) values for all possible overhang sequences for a complete set of all possible sequence combinations of overhangs with a single length under selected experimental conditions. The first database may be the product of an analysis of annealed overhangs, examples of which are provided in Figures 2A-2C and Example 1. The second database may be derived using the same assay to provide data on the frequency and fidelity of ligation by various ligases that recognize different 4-base overhangs and have different or similar biases. The complete set of overhangs may include overhangs of various sizes. Nucleic acids include DNA, RNA, or DNA / RNA hybrids or chimeras. For convenience, DNA may be detailed in the description, examples, and claims of this specification, but the embodiments of this specification are not limited to DNA and may be applied to any type of nucleic acid as described above.
[0128] Factors for determining the appropriate length of the overhang include how many fragments are desired to be ligated, where the longer the overhang, the larger the set of possible combinations. This allows more fragments to be ligated to form the target polynucleotide, each with a unique overhang that is complementary to the overhang of the adjacent fragment. Other factors include melting / annealing efficiency, where shorter overhangs melt and anneal faster and longer overhangs require higher melting temperatures. Ligation efficiency is another factor where longer overhangs may ligate more efficiently than shorter overhangs. Ligation efficiency also depends on the properties of the single or group of nucleotides in the overhang, with some sequences hybridizing and / or ligating to form junctions more efficiently than others, reducing bias and not favoring or inducing mismatches.
[0129] In one embodiment, the output from the system instructs the user which restriction endonuclease to use to cleave the nucleic acid to generate an overhang optimized for ligation fidelity or with a sequence chosen according to a selected ligation fidelity. However, other cleavage enzyme systems such as Uracil-Specific Excision Reagent (USER®, New England Biolabs, Ipswich, MA), Argonaute, Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR), or other cleavage enzymes may be used to generate the overhang.
[0130] The experimental conditions mentioned above are provided to the user from a computer interface by a menu and then selected by the user, or selected by a computer that calculates all the various parameters for assembly and provides the best conditions for efficiently correctly joining all the fragments in the set. The use of type IIS restriction enzymes allows for precise selection of the site where the DNA is broken, allowing for the elimination of the restriction enzyme recognition sequence from the final construct (thus allowing for a seamless one tube assembly reaction) or for the elimination of certain nucleic acid assemblies. For example, for gene coding regions, intact junctions that do not change the DNA sequence are important. In other applications, such as cistron formation, additional or modified nucleotides that may remain from the assembly reaction may not interfere with gene expression of the target nucleic acid. In one embodiment, endonucleases suitable for use in the generation of overhangs and intact junctions include: 2-base overhang generators (e.g., BtsI and its isoschizomers, AcuI and its isoschizomers), 3-base overhang generators (e.g., SapI and its isoschizomers, and BspQI and its isoschizomers (both recognize 7 bases)); 4-base overhang generators (e.g., BsaI-HFv2 and its isoschizomers (6-base recognition), BbsI and its isoschizomers (6-base recognition), BsmBI and its isoschizomers (6-base recognition), PaqCI (7-base recognition), and A 5-base overhang generator (e.g., HgaI and its isoschizomers, which have a 5-base recognition site).
[0131] Other restriction endonucleases and their isoschizomers described in the New England Biolabs 2017 / 2018 catalog can be used for those assembly reactions that do not need to be intact.
[0132] A 2-base overhang will generate a 16x16 matrix data table, a 3-base overhang will generate a 64x64 matrix data table, a 4-base overhang will generate a 256x256 matrix data table, a 5-base overhang will generate a 1024x1024 matrix data table, and a 6-base overhang will generate a 4096x4096 matrix data table. The upper limit for the length of an overhang using a type IIS restriction endonuclease can be 5, 6, 7, or 8 bases long. For nicking agents such as USER, the number of bases in the overhang can be as many as the user desires based on the positioning of uracils. The set of optimized Watson-Crick pair overhangs contains overhang pairs that can efficiently ligate with the correct complementary partner, are not palindromic, and are unique within the set. Other overhang pairs are preferably tolerated as long as no individual overhang forms a ligation product with an overhang partner that contains one or more mismatches, preferably no more than one mismatch. The highest fidelity set of overhangs with good ligation fidelity can be provided by the computer for any selected number of junctions (e.g., 10 junctions, 12 junctions, 15 junctions, 20 junctions, etc.). The higher the number of junctions, the lower the average maximum ligation fidelity of the set of overhang pairs.
[0133] In some embodiments, overhangs are generated using alternative enzymes, such as nicking agents, such as USER (see also, e.g., US 7,435,572), or EndoMS, which are suitable for generating overhangs in DNA fragments, and Argonaute and Cas cleavage enzymes, which are suitable for overhangs in DNA and RNA, and which utilize a guide DNA or RNA.
[0134] Embodiments of the method allow a user to receive computational output that provides an optimized set of overhangs based on measures of the net effect of cleavage, melting, annealing, and ligation for a particular combination of cleavage enzyme and one or more ligases, under a given set of cycling conditions, some or all of these characteristics provided by the user. The output can then provide a relative ligation efficiency and / or ligation fidelity value for each overhang pairing.
[0135] The computational output may additionally provide the user with an optimized protocol for performing the assembly to obtain the desired overall ligation fidelity detailing at least one of the temperature, time of hybridization, cycling conditions for ligation, and buffers.
[0136] The computational output may include a graphical feature output including one or more of the following: (1) the entire assembled sequence with junction sites highlighted, (2) a map of the input fragments with individual cleavage sites indicated on the fragments, where the set of cleavage sites has been computationally determined to obtain an optimal set of overhangs for fragment assembly to form the desired product, (3) a matrix of ligation fidelity of the selected overhangs under user-specified or computer-optimized experimental conditions, and (4) a set of primer sequencers including the selected Type IIS restriction endonuclease recognition sequence and overhang sequence plus any additional target fragment sequences to direct automated oligonucleotide synthesis. The set of primer sequences may be electronically transferred to a receiving location to direct a DNA synthesizer to generate such primers.
[0137] In one embodiment, the results of a user-selected set of overhangs can be optimized by providing a set of conditions that the user prefers to achieve efficient and accurate hybridization. Short linkers of any sequence are preferred for a large number of fragments (e.g., ≧20). Multiple data sets providing overhang optima under various conditions may be accessed. Such an assay allows the user to select a set of enzymes and reaction conditions that give the highest possible fidelity and efficiency for the selected set of overhangs.
[0138] In another embodiment, partial overhang pair reaction parameters and data sets may be selected by the user, and partial overhang reaction parameters and optionally data sets may be selected by the computer to provide the optimal ligation efficiency and fidelity possible to create the desired number of ligation fragments. For example, a total of 15 junction pairs may be required to link 16 fragments of double-stranded nucleic acid fragments. Here, six overhang pairs are selected by the user, and the remaining overhangs are provided in the computer-generated output, optionally with preferred experimental conditions including the selection of a ligase. The user can then receive an additional 9 optimized overhang pairs with optional selection of reaction components such as restriction enzymes, ligases, and other reaction conditions, such as optional selection of cycling time and temperature, that provide the highest possible ligation fidelity and efficiency for the final set of 15 members.
[0139] In another embodiment, a user inputs a gene, gene pathway, plasmid or chromosome sequence into a computer for splitting into fragments suitable for high fidelity and efficient assembly using an optimized set of overhangs. The user may specify the target nucleic acid and the desired number of fragments. A web tool or graphical interface provides sequences of the desired number of fragments at optimal junctions that meet the hybridization parameters of the relevant overhangs that form intact junctions when ligated, thus allowing the user to generate the target polynucleotide in the desired manner. If the user additionally specifies a minimum acceptable fidelity, the sequence specifications for the desired number of fragments may be modified, and in fact the number of fragments provided to the user may be modified to provide the maximum number of sequences possible at junctions that provide the specified minimum acceptable fidelity.
[0140] In another embodiment, the user may provide a target sequence and may further specify some junctions and some subset of reaction conditions (or all reaction conditions) to be included in the design of the building blocks with a given overhang, and the computer provides the user with a list of overhangs for efficient ligation that provide the best additional junctions and / or reaction conditions.
[0141] Assembly proceeds at a single temperature appropriate for either all types of enzymatic activity (e.g., cleavage and ligation enzymes) used in the reaction or any number of cycling conditions varying between optimal cleavage / melting and optimal annealing / ligation temperatures. That way, overhangs can be generated and sealed in one pot, and multiple fragments can be ligated together in one experiment.
[0142] Implementation of nucleic acid assembly using computer programs and general-purpose computer systems The various components of the various systems described herein may be implemented as computer programs using a general purpose computer system. Such a computer system typically includes a main unit connected to both output devices for displaying information to a user and input devices for receiving input from the user. The main unit generally includes a processor connected to a memory system via an interconnection mechanism. The input and output devices are also connected to the processor and memory system via the interconnection mechanism.
[0143] One or more output devices may be connected to the computer system. Exemplary output devices include, but are not limited to, liquid crystal displays (LCDs), plasma displays, cathode ray tubes, video projection systems, and other video output devices, printers, and devices for communicating over low or high bandwidth networks, including network interface devices, cable modems, and storage devices such as disks or tapes. One or more input devices may be connected to the computer system. Exemplary input devices include, but are not limited to, keyboards, keypads, trackballs, mice, pens, tablets, touch screens, cameras, communication devices, and data input devices. The invention is not limited to the particular input or output devices used in combination with the computer system or the input or output devices described herein.
[0144] The computer system may be a general-purpose computer system programmable using a computer programming language, a scripting language, or an assembly language. The computer system may also be specially programmed, special purpose hardware. In a general-purpose computer system, the processor is typically a commercially available processor. A general-purpose computer also typically has an operating system that controls the execution of other computer programs and provides scheduling, debugging, input / output control, accounting, compilation, storage allocation, data management and memory management, and communication control and related services. The computer system may be connected to a local network and / or a wide area network such as the Internet. The connected network may transfer to and from computer system program instructions for execution on the computer, media data such as video data, still image data, or audio data, metadata, review and approval information for media compositions, media annotations, and other data.
[0145] A memory system typically includes a computer readable medium. The medium may be volatile or non-volatile, writable or non-writable, and / or rewritable or non-rewritable. A memory system typically stores data in binary format. Such data may define an application program executed by a microprocessor, or information stored on a disk that is processed by the application program. The invention is not limited to a particular memory system. Time-based media may be stored on or input from magnetic, optical, or solid-state drives, which may include an array of local or network-attached disks.
[0146] Systems such as those described herein may be implemented in software, hardware, firmware, or a combination of the three. The various elements of the system, individually or in combination, may be implemented as one or more computer program products in which computer program instructions are stored on a computer-readable medium for execution by a computer or transferred to a computer system over a connected local or wide area network. Various steps of a process may be performed by a computer executing such computer program instructions. The computer system may be a multi-processor computer system or may include multiple computers connected over a computer network. The components described herein may be separate modules of a computer program or may be separate computer programs operable on separate computers. Data generated by these components may be stored in a memory storage system or transmitted between computer systems by various communication media, such as carrier signals.
[0147] Uses of Polynucleotide Ordered Assembly The improved methods, compositions and kits may be used in a number of diagnostic and medical contexts, some examples of which are provided below.
[0148] Example 5 describes the use of multi-fragment assembly methods for coronavirus component sequences that can be engineered into transcription of novel virion sequences into RNA and tested as a potential basis for vaccine development. Figures 14 and 15 show the improved multi-fragment assembly described herein for the design of CarT cell therapy and phage antibiotics. These methods rely on creating large arrays of 25-50 fragments in a one-pot reaction. These arrays may then be used to target multiple genes in a single one-pot reaction. Alternatively, a single gene may be targeted multiple times and / or at multiple locations to increase the efficiency of editing via homology-directed repair. Another use of arrays is to create or engineer genetic pathways, for example, for metabolic pathways. This metabolic pathway may then be controlled by dcas9 and coupled to those genes that are desired to be turned off. Biosensor arrays may be designed (e.g., using Cas13) to sense a wide range of nucleic acids at once, for example, in a multipathogen sensor system.
[0149] In the applications listed above, the overall design rules provided by the computer and summarized in Figures 9A-9D facilitate the creation of 25-50 fragment arrays almost instantly. It is possible to use computer tools to create arrays of fragments with relatively low fidelity (see hlig3). hlig3 facilitates the creation of libraries in which the number of gRNA template sequences included is varied by controlled deletion via mismatch ligation.
[0150] For arrays utilizing small fragments (e.g., 30-70 nucleotides), these can be easily synthesized by an oligonucleotide foundry using information from computational tools about the overhangs that work best for subsequent assembly that relies solely on ligation. Designers can select the optimal ligase and preferred overhang length for the optimal frequency of correctly assembled fragments and the desired fidelity of the assembled fragments in a single-step reaction.
[0151] All publications, patents, and patent applications mentioned in this specification are incorporated by reference into this specification as if each individual publication, patent or patent application was specifically and individually indicated to be incorporated by reference, including U.S. Provisional Patent Application No. 63 / 125,530 (filed December 15, 2020); U.S. Provisional Patent Application No. 63 / 213,807 (filed June 23, 2021); and U.S. Provisional Patent Application No. 63 / 213,859 (filed June 23, 2021). EXAMPLES
[0152] [Example 1] Differences in ligation frequency for different ligases caused by different 4-base sequences All enzymes (except hLig3) and buffers were obtained from New England Biolabs (NEB, Ipswich, Mass.). T4 DNA Ligase Reaction Buffer (1x) is 50 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 1 mM ATP, 10 mM DTT. NEBNext® Quick Ligation Reaction Buffer (1x) is 66 mM Tris pH 7.6, 10 mM MgCl2, 1 mM DTT, 1 mM ATP, 6% polyethylene glycol (PEG6000) at 25°C. NEBuffer2 (1x) is 10 mM Tris-HCl (pH 7.9), 50 mM NaCl, 10 mM MgCl2, 1 mM DTT. CutSmart® Buffer (1×) is 20 mM Tris-Acetate (pH 7.9), 50 mM Potassium Acetate, 10 mM Magnesium Acetate, 100 μg / ml BSA. ThermoPol® Buffer is 20 mM Tris-HCl (pH 8.8), 10 mM (NH4)2SO4, 10 mM KCl, 2 mM MgSO4, 0.1% Triton-X-100. Standard Taq polymerase buffer is 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2. All column cleanups of oligonucleotides and ligation libraries were performed using Monarch® PCR & DNA Cleanup Kit columns (New England Biolabs, Ipswich, MA) following the oligonucleotide cleanup protocol. Oligonucleotide purity and sizing were performed using an Agilent Bioanalyzer® 2100 (Agilent, Santa Clara Calif.) using the DNA1000 assay according to standard protocols.
[0153] The hLig3beta gene was synthesized by Biomatik (Ontario, Canada) and subcloned into the pET28 plasmid in frame with an N-terminal His6 tag. The construct was cloned into the T7 Express lysY / I qExpression was performed in E. coli cells (New England Biolabs, Ipswich, MA).
[0154] Substrates for the 4-base overhang ligation fidelity assay were generated according to WO2020 / 081768 and Potapov et al. (2018) ACS Synthetic Biology, 7, 2665-2674. Briefly, the initial PAGE-purified substrate precursor oligonucleotides contained a 5'-end region, a randomized 4-base region, a BsaI-HFv2 binding site, a constant region, an internal 6-base randomized region as a control for synthesis bias, and a region corresponding to the SMRT Bell sequencing adaptor for Pacific Biosciences SMRT sequencing. The precursor oligonucleotides were extended as previously described and purified using the Monarch PCR & DNA Cleanup Kit. The extended DNA was cleaved using BsaI-HFv2 to generate 4-base overhangs.
[0155] For each ligation reaction, substrate (100 nM) was combined with DNA ligase (either T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, PBCV-1 DNA ligase, or hLig3 at a final concentration of 1.75 μM) in 1× T4 DNA ligase buffer (or NEBNext® Quick Ligation reaction buffer for reactions marked with PEG) in a total reaction volume of 50 μL and incubated at 25° C. for 1 hour. Reactions were quenched with 2.5 μL of ligase reaction quench (500 mM EDTA + 2.5% v / v Proteinase K) and samples were heated to 37° C. for 30 minutes to allow Proteinase K DNA ligase cleavage. Reactions were then purified using the Monarch PCR & DNA Cleanup Kit following the oligonucleotide cleanup protocol. Each ligation was performed with minimal replicates, and ligation yields were determined by Agilent Bioanalyzer (DNA1000), with errors reported as one standard deviation. Ligation libraries were treated with Exonuclease III (50 U) and Exonuclease VII (5 U) in a 50 μL volume of 1× standard Taq polymerase buffer, incubated for 1 h at 37 °C. Libraries were purified using the Monarch PCR & DNA Cleanup kit, oligonucleotide cleanup protocol including a second wash step, and then quantified by Agilent Bioanalyzer (DNA1000). Typical concentrations of the final libraries were between 0.5 and 2 ng / μL. Two replicates were performed for each ligase. Sequencing and analysis of sequencing data were performed as previously described in WO2020 / 081768 and Potapov et al. (2018) Nucleic Acids Research, 46, e79-e79. A consensus sequence of the top and bottom strands of the ligation products was generated and the actual overhang sequence on each strand was extracted. The frequency of all overhang pairs observed in the ligation products was tallied and used to derive results. Subsequent analysis was performed after combining the data from the replicates.Complete results from each experiment are shown in Figures 3Ai / 3Aii-3Hi / 3Hii.
[0156] To determine the fidelity and bias profiles of DNA ligases in end ligation, sequencing libraries were prepared by mixing each DNA ligase (T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, PBCV-1 DNA ligase, and hLig3) with DNA hairpin substrates containing degenerate 5'-4 base overhang ends, allowing for each ligase to observe every possible sequence context in a single reaction (Potapov et al. (2018) Nucleic Acids Research, 46, e79-e79). Ligases were present in large excess compared to the DNA substrate to allow rapid ligation of short sticky ends. Following the ligation reaction, libraries were sequenced using PacBio SMRT sequencing. A summary of the multiplex ligation data for each ligase, including the total number of ligation events, the percentage of correct (Watson-Crick) vs. incorrect (mismatch) ligations, and the yield of the resulting ligation products, is shown in Table 2.
[0157] [Table 2]
[0158] Since the population of DNA hairpin substrates presented a complex equilibrium system in which ligation had to find compatible ends, it was expected that the presence of competing annealing partners would limit ligation yields due to the presence of annealed pairs that ligated with low efficiency. The library ligation yields at 1 hour differed significantly among the ligases tested. T4 DNA ligase, T3 DNA ligase, and hLig3 all yielded more than 55% ligation product and are generally among the most efficient end-joining ligases. PBCV-1 ligase had a somewhat lower yield (50%), reflecting a less robust end-joining activity. T7 DNA ligase had a much lower yield, reaching only 20% ligation product.
[0159] The multiplex ligation data revealed ligation sequence bias in favorable overhang sequences. In this assay, the number of reads for each overhang was a proxy for its ligation efficiency; sequence bias for each ligase was estimated from the relative frequency of each overhang appearing across all ligation products. Variable overall bias and disparate preferred sequences were detected between ligases (Figure 4A). T7 DNA ligase showed the highest degree of sequence bias. All other ligases tested had a much closer distribution of ligation frequencies, but there were differences in how tightly the data points were clustered around the mean. Both T4 DNA ligase and hLig3 showed the least amount of bias, ranging from values less than 2-fold compared to T7 DNA ligase. PBCV-1 and T3 had similar average ligation frequencies, but a slightly larger range of observed ligation frequencies.
[0160] When analyzing the ligation frequency of individual overhangs, the preferred or unpreferred specific sequences varied between enzymes. For most ligases, a weak overall tendency was observed placing overhangs with higher GC content in more efficient ligation (Figure 4A). A bias in favor of high GC pairings was seen for both Watson-Crick ligation and pairings containing at least one mismatch, showing a preference for more strongly annealed sequences in both cases. T7 DNA ligase was the most vulnerable to this bias, with low GC overhangs (0% or 25% GC content) rarely ligated, and high GC content (>50%) accounting for 96% of the ligated products. These data indicate that for T7 DNA ligase, end-joining ligation efficiency is governed by the GC content of the overhang. For T4 DNA ligase, T3 DNA ligase, and PBCV-1, a dependence on GC content could still be observed, although it was less pronounced. Conversely, hLig3 ligation appears to be independent of GC content. However, clearly GC content is not the only factor contributing to the difference in bias. Although experimental replicates of each individual enzyme are consistent in their preferred overhang sequences, comparison of sequence preferences between different DNA ligases reveals additive and complex differences that cannot be easily explained by GC content or other simple trends.
[0161] The majority of correctly base-paired ligation partners were observed at similar overall frequencies for T4 DNA ligase, T3 DNA ligase, PBCV-1 ligase, and hLig3. Overhangs with TNNA sequences ligated inefficiently and were reduced compared to the median (Table 2). The corresponding ANNT overhangs did not show reduced occurrence compared to other overhangs in the set, despite being predicted to be present in the same proportions in the substrate pool. The data indicated that there is an underlying inefficiency in ligating overhang pairs that both contain 5'-T. Furthermore, overhangs containing 5'-C were ligated with significantly reduced efficiency by hlig3 (Figure 3Aii).
[0162] The ligases examined here showed extreme fidelity with T7 DNA ligase showing the highest fidelity (89% correct ligation products), whereas hLig3 had the lowest fidelity (56% correct ligation products). T4 DNA ligase showed moderate fidelity (72% correct ligation products). T4 DNA ligase, T3 DNA ligase, PBCV-1 ligase and hLig3 had a wide range of fidelity for individual overhang sequences, with some overhangs showing very few mismatch ligation events and others showing frequent mismatch ligation (Figure 4b). For many overhangs, ligation products were almost exclusively Watson-Crick partners, even when all possible partners were presented. Where specific mismatch base pairs tolerated by each ligase occurred, these were more prevalent at the 5'-terminal nucleotides ("end") rather than in the middle of the overhang ("middle"). Although some frequent mismatches, especially G:T mismatches, were common among all tested ligases, differential mismatch pairings were also observed among the ligation products of each ligase.
[0163] For example, when T4 DNA ligase was presented with all possible ligation partners, some overhangs paired with Watson-Crick partners in more than 90% of the ligation products (e.g., AAAA, AAGA, ACAA, GAAA). Other overhangs ligated to partners containing at least one mismatch; some overhangs paired with partners containing mismatches more than 60% of the time (e.g., GGCG, GGCC, GGGC, GGGG).
[0164] T4 DNA ligase, with an overall fidelity of 72%, had a median fidelity of 90% for overhangs with 0% GC content, with the average fidelity decreasing with each incremental increase in GC content, eventually dropping to 52% fidelity for overhangs with 100% GC content (Figure 5A). Of all ligation products containing an overall 28% mismatch, 98% of the ligation products had only a single mismatch. Mismatch ligations at the end position (N1) of the 4-base overhang were dominated by G:T and T:G mismatches, accounting for 65% of all mismatch ligations at the ends. The presence of mismatches at the middle positions (N2 and N3) of the overhang was less tolerated by T4 DNA ligase, but still dominated by G:T mismatches.
[0165] In contrast, hLig3 showed a wide range of ligation fidelity. Most overhangs ligated with less than 50% fidelity, and some overhangs (TAAG, AATA, TTAC, CCAA) ligated with more than 80% fidelity. The effect of GC content was weaker for hLig3, with an average fidelity of 72% for overhangs with 0% GC content and 32% for overhangs with 100% GC content (Figure 5C). More than half (56%) of the ligation products contain mismatched base pairs. hLig3 has a significant accumulation of mismatch products with more than a single base pair mismatch, with 8% of the ligation products containing two mismatches. Of these double mismatches, the majority (97%) contain at least one mismatch at the end position, typically containing at least one G:T mismatch. In addition, although G:T and T:G mismatches were well tolerated, hLig3, T3 DNA ligase, and PBCV-1 ligase were also more tolerant of purine:purine mismatches at both end and internal positions, with G:A and G:G mismatches ligated approximately as frequently as G:T mismatches.
[0166] In contrast, T7 DNA ligase had a narrower range of ligation fidelity, with only a small number of overhangs ligated with fidelity below 80%. T7 DNA ligase showed an average fidelity of over 86%, regardless of GC content. T7 DNA ligase had an overall low tolerance for mismatch ligation, with only 12% of ligation products containing mismatches. Similar to T4 DNA ligase, single base pair mismatches accounted for nearly all (98%) of T7 DNA ligase mismatch ligation products, with the predominant mismatches being G:T and T:G at the end positions, and G:T at the middle position of the 4-base sequence.
[0167] [Example 2] Bioinformatic tools for designing Golden Gate assemblies The computational design tool for determining overhangs to optimize ligation fidelity in FIG. 9A has three components that together form the ligation fidelity tool: the ligase fidelity viewer (see FIG. 9B), the GetSet viewer (see FIG. 9C), and the SplitSet interface (see FIG. 9D) (see, for example, WO2020 / 081768). All three computational design tools rely on a single ligase (T4 DNA ligase). The data obtained here is added to these three tools by providing a preference selection under the ligation conditions menu. Ligases with different base sequence preferences that influence the overhang selection can be selected. The benefit of this additional data improves the accuracy of the tool for the ordered assembly of multiple fragments. The data is obtained from 4-base overhangs, but can easily be repeated for 2-, 3-, and 5-base overhangs.
[0168] The data also provides the user with a refined estimate of assembly fidelity for a given set of user-supplied overhangs, and identifies problematic overhang pairings where this is undesirable, with a high probability of mismatch ligation.
[0169] The GetSet tool allows users to generate overhang sets with maximum assembly fidelity using automated overhang selection. GetSet returns high-fidelity overhang sets that match the input criteria of number of overhangs, overhang length, and ligation conditions. Users can specify overhang sequences that should be included or excluded from the results. Importantly, GetSet does not use pre-computed results, but instead uses a probabilistic search algorithm to identify de novo high-fidelity overhang sets. As a result, the probabilistic search algorithm can return different recommended overhang sets from the same input criteria, i.e., repeating the search may result in different junctions with similar predicted fidelity. Therefore, the ability to save and recall previous GetSet search results is included. As an example, the GetSet tool was used to extend a standard overhang set used in plant synthetic biology: the set size could be increased from 11 overhangs to 20 overhangs, with only a slight decrease in predicted assembly fidelity from 81% to 80%.
[0170] The SplitSet tool designs high-fidelity assembly fragments from a desired target DNA sequence. To use this tool, users input the DNA sequence, the desired number of fragments, ligation conditions, and an approximate search window for fusion sites (by default, the program selects an equal-spaced search interval). The SplitSet tool splits the input DNA sequence at the highest-fidelity junction set within the selected parameters. Additionally, users can exclude specific fusion site sequences to ensure compatibility with existing modular cloning systems or include fixed sites by setting a narrow search window to encompass the sites that need to be used.
[0171] Additional features include checking fragments for the presence of any internal sites that may affect the choice of type IIS restriction enzyme to direct the assembly, or alerting the user to remove such internal sites via domestication. The program can also automatically generate a set of primers for the DNA fragments to add the necessary flanking bases and recognition sites either for amplicon generation of the insert to be used directly or for precloning purposes. Finally, a report can be generated describing the entire assembly, including a color-coded graphical readout, the final assembly sequence, and a description of each junction between the inserts.
[0172] [Example 3] Aprataxin / 5'-deadenylase and PEG in Golden Gate assembly In the reactions described in the tables set forth below, the ordered assembly reactions consisted of precloned DNA fragments (3 nM each) or PCR fragments (3 nM each) and the indicated amounts of restriction enzyme and ligase in T4 DNA ligase buffer or Blunt T / A ligase buffer (New England Biolabs, Ipswich, Mass.). Assembly reactions were incubated as described in Example 3 using one of three different ligation protocols, followed by a final incubation step of 5 minutes at 60° C. and then a final hold at 4° C. before transformation. Transformations were performed using 2 μl of each assembly reaction added to 50 μl of competent T7 Express chemically competent E. coli (New England Biolabs, Ipswich, Mass.) plated on the appropriate growth medium and incubated at 37° C. to allow colonies to form.
[0173] The correctly assembled target polynucleotide encoded a lac operon cassette (approximately 5 kb), allowing blue colonies of correctly assembled fragments to be distinguished from white colonies containing incorrect assemblies on IPTG / Xgal / chloramphenicol plates.
[0174] In general, the addition of 5' deadenylase enzyme or PEG600-6000 was found to increase the yield of correct assembly.
[0175] [Table 3]
[0176] [Table 4]
[0177] [Table 5]
[0178] [Table 6]
[0179] [Example 4] Examination of the effect of reaction temperature on multifragment assembly fidelity and assembly of a lac operon cassette from 52 fragments. Multifragment assemblies that rely on a two-enzyme mixture (restriction endonucleases and ligases) typically use a two-step cycling protocol that alternates between a 16°C incubation step to maximize DNA ligation efficiency and a 37-42°C incubation step to maximize fragment digestion efficiency. Because higher reaction temperatures have been shown to improve DNA ligase fidelity, omission of the 16°C incubation was tested to determine its impact on the fidelity of multifragment assembly. The frequency of multifragment assembly errors at 37°C or 42°C was quantified with a multiplexed high-throughput DNA sequencing assay, and the results were compared to reactions using conventional thermal cycling protocols of 37 / 16°C or 42 / 16°C. Reactions performed at constant incubation temperatures were incubated for an extended period of 16 hours to compensate for the reduced ligation efficiency. The results showed that when the 16° C. incubation step was omitted, the frequency of ligation errors was reduced by more than two-fold, and all mismatch pairs appeared less frequently.
[0180] Multi-fragment assembly reactions could exceed 50 fragments per reaction using a single temperature incubation (static). In this example using BsaI or BsmBI restriction endonucleases, the estimated assembly fidelity dropped to less than 10% at 50 fragments for conventional 37 / 16°C or 42 / 16°C cycling conditions, whereas the 37°C or 42°C static incubation protocol allowed more than 50 fragments to be assembled with greater than 40% accuracy. At static 37°C or 42°C static temperatures, longer incubation times were used to compensate for the loss in efficiency caused by omitting the 16°C incubation step.
[0181] The mismatch frequencies of the assembly reactions were grouped according to nucleotide mispairs (A:A, A:C, A:G, C:C, C:T, G:G, G:T, T:T). Assembly reactions were performed with T4 DNA ligase and either BsaI-HFv2 at 37°C or BsmBI-v2 at 42°C. For comparison, the mismatch frequencies of assembly reactions using conventional thermal cycling protocols with T4 DNA ligase and either BsaI-HFv2 at 37°C and 16°C or BsmBI-v2 at 42°C and 16°C are shown. The mismatch frequencies using BsaI-HF-v2 (37°C) or BsmBI-v2 (42°C) at a single temperature were significantly lower than observed with cycling.
[0182] To test the predictions made herein, a 4.9 kb cassette of the lac operon was cloned into an E. coli destination vector from 52 component parts in a single assembly round. Importantly, the lac operon cassette system used here mimics a traditional cloning reaction, and when the assembly reaction is transformed into E. coli cells, colonies carrying correctly or incorrectly assembled constructs can be easily observed. This test system provides a colorimetric readout to distinguish between transformants carrying correctly and incorrectly assembled products.
[0183] Clonogenic Assay Assembly reactions to reconstruct the lac operon cassette were transformed into chemically competent E. coli cells and colony-forming units were scored as carrying correctly or incorrectly assembled constructs by blue-white reverse display as described above. Briefly, transformations were performed according to the manufacturer's instructions using 2 μL of each assembly reaction added to 50 μL of T7 expression competent cells. The resulting growth was plated on agar plates (Luria-Bertani broth supplemented with 1 mg / mL dextrose, 1 mg / mL MgCl2, 30 μg / mL chloramphenicol, 200 μM IPTG and 80 μg / mL X-gal). Importantly, transformants carrying correctly assembled constructs turn blue after incubation in medium containing IPTG and X-Gal, whereas transformants carrying constructs with assembly errors form white colonies.
[0184] Verification of the 52-fragment lac operon cassette assembly Plasmid DNA was isolated from 18 blue colonies using the Monarch Plasmid Miniprep kit (New England Biolabs, Ipswich, MA). Twelve of the resulting constructs were subjected to PCR with amplification primers flanking the desired insertion site. All constructs yielded amplicon sizes consistent with assembly of all 52 fragments, indicating that the blue colonies contained the desired number of inserts. Six of the isolated constructs were sequenced using nine different sequencing primers spanning the entire predicted insert of 4.9 kb. All six constructs contained regular, error-free assembly of all 52 inserts.
[0185] 52 fragment Lac assembly: 4 base overhang (BsaI-HFv2): Protrusion:
[0186] [ka] (Fragment sizes (base pairs): 83, 125, 103, 45, 114, 83, 122, 81, 89, 139, 81, 119, 62, 96, 92, 92, 107, 113, 85, 82, 113, 95, 110, 78, 96, 78, 101, 103, 115, 75, 113, 98, 115, 63, 99, 126, 93, 69, 102, 86, 139, 56, 121, 73, 128, 110, 67, 93, 144, 90, 96, 77)
[0187] 49% of the observed transformants were found to harbor correctly assembled constructs, in close agreement with the predicted fidelity of 40%. To further confirm successful assembly of all 52 inserts, constructs were purified from a subset of colonies and inserts were analyzed by PCR and Sanger sequencing: all colonies subjected to additional screening were found to harbor constructs with inserts of the expected size and sequence.
[0188] Plasmid DNA was isolated from 18 blue colonies using the Monarch Plasmid Miniprep kit (New England Biolabs, Ipswich, MA). Twelve of the resulting constructs were subjected to PCR with amplification primers flanking the desired insertion site. All constructs yielded amplicon sizes consistent with assembly of all 52 fragments, indicating that the blue colonies contained the desired number of inserts. Six of the isolated constructs were sequenced using nine different sequencing primers spanning the entire predicted insert of 4.9 kb. All six constructs contained regular, error-free assembly of all 52 inserts.
[0189] Taken together, we demonstrate that one-step assembly of phage T7 DNA and lac operon cassettes is an efficient and cost-effective means to create and engineer variants of large / complex DNA constructs that are difficult to obtain and manipulate by current cloning and gene synthesis methods. Here, multi-fragment assembly is shown for the rapid assembly of toxic and / or high molecular weight DNA constructs from dozens of smaller component parts that are easily manipulated and propagated using standard molecular biology techniques. These findings dramatically reduce the number of hierarchical assembly rounds required to generate large constructs by in vitro assembly and can be utilized to assemble entire metabolic pathways, or even small genomes, in a single reaction.
[0190] [Example 5] Rapid one-pot DNA molecule construction from 50 fragments of 40Kb T7 phage DNA. Enzymes, buffers, and media were obtained from New England Biolabs, Ipswich, MA (NEB, Ipswich, MA) unless otherwise stated. Synthetic oligonucleotides were obtained from either Integrated DNA Technologies (IDT, Coralville, IA) or Sigma Aldrich (Sigma, St. Louis, MO). Because the phage genome contains many genes that are toxic to E. coli cells, phage gDNA was reconstructed from PCR-generated DNA fragments to avoid subcloning of toxic genes. Using this strategy, 16 silent mutations were introduced into the phage genome to remove pre-existing BsmBI type IIS restriction sites in the genome. These modifications served the dual purpose of enabling type IIS assembly and serving as marker mutations for assembly verification.
[0191] Assembly of a large target from many small fragments in a single round was tested as follows: 52 fragments were used to construct a 38 kb T7 bacteriophage genome. 52 different optimized 4-base overhangs were selected from the intact T7 viral genome by SplitSet, and the set of selected 4-base overhangs was validated by GetSet using the ligation conditions in the drop-down menu for the conditions specified below.
[0192] Overhang sequence obtained by data-optimized assembly design 52 fragment T7 phage assembly (BsmBI) Protrusion:
[0193] [ka] (Fragment sizes (base pairs): 779, 918, 512, 465, 810, 756, 731, 820, 690, 759, 813, 743, 644, 1043, 494, 887, 526, 918, 854, 837, 540, 880, 678, 812, 732, 802, 976, 530, 745, 568, 885, 1130, 148, 1008, 847, 667, 748, 831, 842, 753, 947, 428, 928, 411, 301, 1383, 424, 1056, 893, 653, 815, 1735)
[0194] Multi-fragment assembly reactions Assembly fragments were generated by PCR (Q5® Hot-Start High-Fidelity 2X Master Mix (New England Biolabs, Ipswich, MA)) using oligonucleotide primers (IDT) and purified using the Monarch PCR & DNA Clean-up Kit. Fragment quality was assessed using an Agilent Bioanalyzer 2100, and each assembly part was quantified using the Qubit Assay (Thermo Fisher Scientific, Waltham, MA). Multi-fragment assembly reactions (5 μL final volume) were carried out with 3 nM of each DNA fragment and 0.5 μL of the appropriate multi-fragment assembly mix (NEB® Golden Gate Assembly Mix (New England Biolabs, Ipswich, MA) in 1× T4 DNA ligase buffer. The BsmBI-v2 mix was used to assemble the T7 phage genome. The reaction generating the T7 bacteriophage genome was cycled between 42° C. and 16° C. for 5 min at each temperature for 96 cycles, followed by incubation at 60° C. for 5 min and a final hold at 4° C. prior to transformation into E. coli.
[0195] Plaque assay The assembled T7 phage genome was transformed into NEB 10 beta electrocompetent cells according to the manufacturer's instructions using 1 μL of reaction mixture in 25 μL of competent cells. The transfection mixture was recovered in 975 μL of NEB 10-beta / stable growth medium and then combined with 3 mL of 50° C. molten top agar (Luria broth containing 0.7% agar). The resulting plates were inverted and incubated at 37° C. for approximately 5 hours until E. coli lawns and phage plaques were visible. Upon transformation, approximately 20 bacteriophage plaques were obtained per ul of assembly reaction, indicating successful assembly of the phage genome.
[0196] Validation of the 52-fragment T7 phage gDNA assembly Several phage plaques were selected for further screening by plaque PCR and restriction enzyme digestion to ensure that they contained a complete and correctly ordered copy of the T7 phage genome. All plaques subjected to further screening contained the predicted genome arrangement and carried the intended silent mutations. Plaque PCR was performed using four sets of amplification primers that together spanned 40 kb of the phage genome. Amplicon lengths were resolved by an Agilient Bioanalyzer 2100 using the DNA 12000 assay. Amplicons from five phage plaques were compared to the parent wt T7 phage genome after restriction enzyme digestion with NdeI or undigested. In both cases, the phage plaques produced patterns identical to the parent wt T7 gDNA. The assembled genome carried the desired silent mutations that removed the native BsmBI restriction site, confirming that the observed plaques were not the result of carryover contamination from parental T7 phage gDNA by digestion of the parental T7 phage genome and plaque amplicons with BsmBI. Digestion of the amplicons from the parental T7 phage genome with BsmBI yielded bands of the predicted size, whereas the amplicons from the assembled phage genome were inactive to cleavage by BsmBI.
[0197] Furthermore, to confirm that the observed phage plaques were the result of in vitro assembly and not the assembly of DNA fragments in E. coli by cellular DNA repair mechanisms, control reactions lacking T4 DNA ligase were performed, and no phage plaques were observed upon transformation of these control reactions. These results demonstrate that the use of high stringency screening allows rapid assembly of over 50 DNA fragments under ordered assembly conditions using a data-optimized assembly design.
[0198] Note that variations in the high-temperature assembly protocol were not necessary to generate infectious phage plaques. We speculate that this is due to the more stringent selection of the plaque formation assay, as improperly assembled variants of the T7 phage genome are less likely to generate viable phage and therefore there is no background from misassembled products. This is in contrast to another successful assembly of a lac operon cassette test system from 52 fragments, where cycling was used. Taken together, these results demonstrate the rapid assembly of phage genomes and suggest that the methods developed here may be applied to other large and / or complex DNA targets.
[0199] [Table 7] TIFF2024500105000011.tif21170
[0200] Overhangs were generated using data with BsaI-HFv2, BsmBI or BspQI.
[0201] The following are the salients used in Table 8:
[0202] Here the overhangs are written in the 5' to 3' direction and the phosphates have been omitted.
[0203] 24 fragment Lac assembly: 4 base overhang (BsaI-HFv2; BsmBI): Protrusion:
[0204] [ka] (Fragment sizes (base pairs): 118, 222, 222, 199, 137, 217, 197, 244, 293, 173, 127, 208, 287, 310, 291, 259, 176, 217, 135, 176, 184, 133, 203, 119)
[0205] 24-fragment T4 phage assembly (BsmBI): Protrusion:
[0206] [ka] (fragment size: 8kb)
[0207] 24-fragment T4 phage assembly (BsmBI): Protrusion:
[0208] [ka] (Fragment sizes (base pairs): 4631, 9473, 7031, 7465, 10181, 1595, 8807, 3851, 4152, 10945, 11254, 7290, 1680, 10213, 7152, 6898, 7162, 7104, 6967, 6827, 8052, 8947, 4304, 7018)
[0209] 30 Pieces Universal Lobe Set (any 4 cutters): Protrusion:
[0210] [ka]
[0211] 52-fragment T7 phage assembly (BsmBI) Protrusion:
[0212] [ka] (Fragment sizes (base pairs): 779, 918, 512, 465, 810, 756, 731, 820, 690, 759, 813, 743, 644, 1043, 494, 887, 526, 918, 854, 837, 540, 880, 678, 812, 732, 802, 976, 530, 745, 568, 885, 1130, 148, 1008, 847, 667, 748, 831, 842, 753, 947, 428, 928, 411, 301, 1383, 424, 1056, 893, 653, 815, 1735)
[0213] 52 fragment Lac assembly: 4 base overhang (BsaI-HFv2): Protrusion:
[0214] [ka] (Fragment sizes (base pairs): 83, 125, 103, 45, 114, 83, 122, 81, 89, 139, 81, 119, 62, 96, 92, 92, 107, 113, 85, 82, 113, 95, 110, 78, 96, 78, 101, 103, 115, 75, 113, 98, 115, 63, 99, 126, 93, 69, 102, 86, 139, 56, 121, 73, 128, 110, 67, 93, 144, 90, 96, 77)
[0215] 100 fragment T4 phage assembly (BsmBI): Protrusion:
[0216] [ka] (Fragment sizes (base pairs): 395, 383, 424, 367, 451, 506, 287, 411, 352, 355, 419, 357, 519, 208, 448, 379, 412, 380, 422, 475, 337, 481, 242, 537, 320, 450, 522, 149, 562, 329, 312, 567, 389, 328, 456, 301, 390, 479, 282, 402, 494, 340, 450, 287, 569, 380, 396, 334, 39 0, 412, 295, 487, 199, 627, 290, 557, 382, 409, 296, 418, 906, 143, 193, 296, 436, 412, 465, 295, 511, 312, 341, 369, 448, 451, 421, 297, 380, 425, 479, 150, 679, 320, 356, 495, 223, 562, 399, 319, 317, 268, 628, 435, 368, 355, 451, 308, 555, 336, 357, 1074)
[0217] [Example 6] Synthesis and genetic engineering of viral genomes for coronavirus research and vaccine development Reverse genetic systems that rely on in vitro ligation provide researchers with a relatively fast / efficient way to generate variants of RNA viruses for research purposes and vaccine development efforts. This methodology has worked to engineer some viruses, but has yet to be successfully applied (despite several attempts) to many high-value targets, including coronaviruses. The large size and inherent virulence / replicative nature of some viral genomes likely precludes maintenance of viral genomic DNA within E. coli cells by conventional techniques. However, efficient assembly of viral genomes from many parts can be achieved using the methods described herein. Virulent viral genomes can be fragmented into small fragments for ease of manipulation within E. coli cells, allowing one-tube assembly of viral cDNA in a matter of hours. As an example, below we demonstrate the design of 12-, 24-, and 50-fragment linear viral genome assemblies (Genbank ID: NC_045512, derived from Severe Acute Respiratory Syndrome Coronavirus 2 isolate Wuhan-Hu-1, complete genome). These fragments may be contained in a plasmid with recognition sequences for selected restriction endonucleases at the insertion sites. For example, the restriction endonuclease may be selected from BsaI-HFv2, BsmBI-v2, BbsI-HF, SapI, BspQI and PaqCI. If PaqCI is selected for virus assembly, an activating molecule as described above should be included in the reaction mixture. The product of cleavage with any of these restriction endonucleases results in an overhang. Optimization of the overhang sequence is provided for various restriction endonucleases according to WO2020 / 081768, with the addition of data for PaqCI (Table 8) and the use of computer tools described in WO2020 / 081768 and incorporated by reference. Reaction conditions for PaqCI cleavage and T4 DNA ligation.Using 75ng / ul of the target plasmid, add 2ul of T4 DNA ligase buffer (10x), 0.5-2ul of PaqCI (10u / ul), 0.25-0.5ul of PaqCI activator (20uM), 0.50-2ul of T4 DNA ligase (400u / ul), and nuclease-free water to 20ul of 1ul for the reaction. The assembly protocol is (37℃, 5min → 16℃, 5min) × 30-60 cycles → 37℃, 5min → 60℃, 5min.
[0218] [Table 8] TIFF2024500105000020.tif216170TIFF2024500105000021.tif219170TIFF2024500105000022.tif216170 TIFF2024500105000023.tif218170TIFF2024500105000024.tif217170TIFF2024500105000025.tif159170
[0219] Segment options for use in assembling coronavirus genomes using multi-fragment assembly methods.
[0220] [Table 9]
[0221] [Table 10] TIFF2024500105000028.tif30161
[0222] [Table 11] TIFF2024500105000030.tif187164
[0223] [Example 7] An automated workflow for generating ordered assemblies of polynucleotides onto a target polynucleotide The workflow could be accomplished on a large scale or entirely on a single machine where the inputs of the various components are presented together or sequentially. In a hypothetical workflow, the desired sequence is input into a computer. The computer then provides an output describing the preferred fragments and overhangs derived from a ligation frequency table to which a set of rules has been added for the ordered assembly of the desired sequence.
[0224] The computer output may interface with a lab-on-a-chip or other equipment containing multiple reagent compartments. Regulation of reaction steps may be controlled on-chip by electrowetting-based liquid transport. For example, the AQdrop® platform (Sharp Life Sciences, Oxford, UK) allows for electronic manipulation of microscale droplets on a "lab-on-a-chip" device. Another platform is acoustic-based liquid transport (Beckman Coulter, Brea, CA). Alternatively, the workflow may be performed using magnetic beads to remove unwanted enzymes / primers from the reaction vessels at different stages as required.
[0225] Thus, the fragments may be synthesized in situ or from a secondary source according to the computer output. The synthesized fragments may be amplified by cloning or by amplification methods such as PCR. The latter may be achieved by combining all the separate synthesized fragments in a single mixture and performing a multiplex PCR. The polymerase may be inactivated and a ligase and a restriction endonuclease may be added to achieve an ordered DNA assembly using the methods described herein. The subsequent assembled target DNA may be (i) incorporated into a vector and then introduced into a host cell by transformation of the vector, (ii) packaged into a virus and introduced into a host cell by infection, (iii) directly introduced into a eukaryotic cell in the form of naked DNA or together with chaperone molecules, or (iv) introduced into an in vitro expression system to determine whether the transcript of the assembled DNA is functional. Optionally, the product of the assembly may be transferred to a platform location to be sequenced, for example by a whole molecule sequencer (Oxford Nanopore or Pacific Biosystems).
[0226] The ordered assembly of DNA molecules using the methods described herein is a powerful tool for synthesizing individual genes or metabolic pathways, and potentially for genetically modifying eukaryotic cells. It also provides a means to synthesize toxic proteins, such as novel nucleases, and determine their specificity and other functions. Ordered assemblies encoding toxic proteins may be transcribed using an in vitro transcription system (New England Biolabs, Ipswich, MA) and then tested for DNA cleavage to determine whether the desired function is achieved. Selected positive proteins may then be produced in cells under specific conditions.
[0227] In one example of an automated workflow in which large molecules are constructed from smaller fragments, the first step is to synthesize a set of fragments at least 20 bases in length enzymatically (e.g., using terminal transferase), or by chemical synthesis, or as PCR products from a larger substrate or a set of overlapping fragments.
[0228] These fragments can be assembled using the protocol described herein. Assembly of 50 fragments of 25 bases produces a target polynucleotide of 1000 bases. The restriction endonucleases and ligases can optionally be heat inactivated at 60° C. prior to the next assembly step.
[0229] The assembly process may be repeated again with the newly generated polynucleotide fragments. For example, primers, aptamers and polymerases for amplifying the newly formed polynucleotide fragments from the previous step may then be generated by multiplex PCR. The amplified polynucleotide fragments are then subjected to restriction endonuclease cleavage and ligation to generate 12,500 bp fragments from 50 x 250 bp polynucleotides, or 50,000 bp polynucleotides from 50 x 1000 bp fragments.
[0230] The assembly may then be repeated, for example, by combining 50 x 12,500 (625 Kb) fragments or 50 x 50,000 bp fragments (2.5 Mb), followed by cleavage and ligation, to generate a 625 kb or 2.5 Mb target polynucleotide.
[0231] Another example of a workflow is one-step DNA assembly using multiple DNA fragments ranging in size from 200 to 1000 bp.
[0232] The efficiency of fragment ligation depends on the overhang, enzyme and experimental conditions, but is independent of the length of the polynucleotide fragments used in the assembly, at least up to 1 kb. Assembly of multiple fragments used in the assembly of more than 24 fragments may be preferentially achieved by extended incubation periods. If these incubation periods exceed 24 hours, it may be preferable to use a static ligation protocol instead of a touchdown or drop-down protocol. wt T4 ligase may be used in thermal cycling up to a temperature defined by the drop-down conditions of 42°C / 16°C. Above 42°C, a thermostable ligase is preferred.
[0233] Suggested Uses of the Workflow Described Herein One tube, multiple constructs: The emulsified ordered assembly workflow allows users to generate a variety of constructs from multiple fragments in a small droplet-based format where "positive" droplets can be sorted for downstream applications by FACS. Mismatch connections during ordered assembly allow users to generate a variety of construct variations in one tube by deliberately using overhang(s) that pair well with multiple partners. For example, users can generate the same gene circuit with several different promoters in one tube and identify the best construct through genetic screening.
[0234] Assembly of higher order DNA structures (using non-standard DNA parts). DNA origami can allow users to assemble DNA structures to facilitate transfection and consistent gene regulation by controlling the shape of the assembled molecules. Generation of branched constructs can allow users to create future constructs with branched configurations for parallel control. For example, the use of non-standard part assembly fragments (such as 1-duplex to 2-duplex connectors) can be placed in coding sequences close to the same insulator element.
[0235] Mixed overhang lengths for ordered assembly resulting from the use of multiple Type IIS enzymes allows for maximizing the number of fragments for assembly, and allows users to increase the number of overhangs possible in a single reaction by mixing orthogonal sets of overhangs of different lengths. This approach can also generate final constructs with bubbles to facilitate downstream applications, such as strand invasion or cleavage of the product by a nuclease (e.g., T7 Endo I).
[0236] While exemplary embodiments have been described herein, it will be apparent to those skilled in the art that the foregoing is merely illustrative and not limiting, presented by way of example only, and numerous modifications and other embodiments are within the purview of those skilled in the art and are contemplated as falling within the scope of the present invention.
Claims
1. A synthetic self-complementary oligonucleotide comprising a double-stranded region and a single-stranded loop, said double-stranded region containing a recognition sequence for PaqCI, having non-ligatable 3' and 5' ends, and not cleavable by PaqCI.
2. 2. The oligonucleotide of claim 1, wherein the double-stranded region is 10 to 50 base pairs in length.
3. The oligonucleotide of any one of claims 1 to 2, which is less than 110 nucleotides in length.
4. The oligonucleotide of any one of claims 1 to 3, wherein the 3' end of the oligonucleotide is not a 3' hydroxyl.
5. The oligonucleotide of any one of claims 1 to 4, wherein the 5' end of the oligonucleotide is not a 5' phosphate.
6. The oligonucleotide according to any one of claims 1 to 5, wherein the recognition sequence is (5'-CACCTGC-3' / 3'-GCAGGTG-5').
7. (a) a synthetic self-complementary oligonucleotide according to any one of claims 1 to 6; and (b) PaqCI, or a variant thereof having an amino acid sequence having at least 90% amino acid sequence identity with SEQ ID NO:
1. A reaction mixture comprising:
8. 8. The reaction mixture of claim 7, wherein the ratio of PaqCI to synthetic self-complementary oligonucleotide ranges from 1 unit of PaqCI:0.75 pmoles to 9 pmoles of oligonucleotide.
9. 9. The reaction mixture of claim 7 or 8, further comprising a double-stranded DNA substrate, said substrate containing a recognition sequence for PaqCI and capable of being cleaved by PaqCI to result in a four-base overhang.
10. 10. The reaction mixture of claim 9, wherein the recognition sequence in the DNA substrate is (5'-CACCTGC-3' / 3'-GCAGGTG-5').
11. The reaction mixture according to any one of claims 7 to 10, further comprising a DNA ligase.
12. 12. The reaction mixture of claim 11, wherein the DNA ligase is selected from the group consisting of T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, PBCV-1 DNA ligase and human ligase 3 (hLig3).
13. 13. The reaction mixture of any of claims 7 to 12, further comprising a plurality of plasmids or PCR products containing fragments each flanked by binding sites for PaqCI, wherein cleavage of said plasmids or PCR products with PaqCI or a variant thereof results in fragments with different four base overhangs.
14. 14. The reaction mixture of any one of claims 7 to 13, wherein the ratio of PaqCI to ligase is 2.5 to 20 PaqCI units to 200 to 800 ligase units.
15. (a) (i) a synthetic oligonucleotide according to any one of claims 1 to 6; (ii) PaqCI; (iii) a ligase; and (iv) a library of DNA substrates each having at least one PaqCI recognition sequence and cleavage site. obtaining a reaction mixture comprising: (b) cleaving the library of DNA substrates with PaqCI to generate fragments with four-base overhangs; and (c) ligating the complementary four base overhangs to each other to produce an ordered assembly of fragments. A method comprising:
16. 16. The method of claim 15, wherein the DNA substrates in the library are selected from one or more of the group consisting of PCR products, plasmids, genomes or chromosomes.
17. The method of claim 15 or 16, wherein (c) further comprises ligating the ordered assembly in a destination vector or a viral genome.
18. The method of claim 17, wherein the destination vector is a plasmid or a chromosome.
19. The method according to any one of claims 15 to 18, wherein the ligase is selected from the group consisting of T4 DNA ligase, T3 DNA ligase, T7 DNA ligase, PBCV-1 and human ligase 3 (hLig3).
20. 20. The method of any of claims 15 to 19, wherein there are 10 to 100 DNA substrates with unique sequences and the ordered assembly comprises 10 to 100 fragments that are ligated together in step (c).
21. 21. The method of claim 20, wherein there are at least 20 DNA substrates with unique sequences and the ordered assembly comprises at least 20 fragments that are ligated together in step (c).
22. 22. The method of any of claims 15 to 21, wherein the reaction mixture further comprises one or more of a DNA repair enzyme, a deadenylase, and / or a crowding agent.
23. 23. The method of claim 22, wherein the crowding agent is polyethylene glycol (PEG) having a molecular weight in the range of 600 to 8000.
24. 23. The method of claim 22, wherein the deadenylase is a yeast deadenylase.
25. The method of claim 22, wherein the DNA repair enzyme is EndoMS.
26. wherein step (a) further comprises identifying a set of four-base overhangs for the reaction mixture using a computer tool; (a) the computer tool generates from the dataset an optimized fidelity and / or frequency score for a set of four base overhangs for a library of DNA, wherein the optimized fidelity and / or frequency score is derived from data on annealing of complementary sequences and data derived from ligase activity on different four base overhangs; and / or (b) a computational tool provides cut points in the in silico sequence to generate fragment sequences for joining in an ordered assembly via optimized four-base overhangs; The method according to any one of claims 15 to 25.
27. A kit comprising the synthetic self-complementary oligonucleotide of claim 1 and PaqCI.
28. 28. The kit of claim 27, further comprising a ligase.
29. 29. The kit of claim 27 or 28, further comprising a cofactor selected from the group consisting of a repair enzyme, a deadenylase and a crowding agent.
30. A kit according to any one of claims 27 to 29, comprising instructions for synthesizing large DNA from component fragments with four base overhangs.
31. A kit according to any of claims 27 to 30, wherein the PaqCI and the activator are combined with the ligase in a single container or are present in separate containers.
32. The kit according to any of claims 27 to 31, wherein at least one of the oligonucleotides, the ligase and the PaqCI variant are lyophilized or immobilized on a solid substrate, such as a two- or three-dimensional surface.
33. The kit according to any one of claims 27 to 32, wherein the crowding agent is polyethylene glycol (PEG) and has a molecular weight in the range of 600 to 8000.
34. The kit according to any one of claims 27 to 33, wherein the repair enzyme comprises a mismatch-specific endonuclease.
35. The reaction mixture of any one of claims 7 to 14, further comprising a DNA repair enzyme, a deadenylase, and / or a crowding agent.
