Ligation and assembly for enzymatic DNA synthesis
30 views
Skip to first unread message
Bryan Bishop
Nov 20, 2022, 2:18:07 PM11/20/22
to enzymatic...@googlegroups.com, Max Berry, Bryan Bishop
Hi all,
Over the past few years, there have been a number of interesting proposals to use ligation for making long oligos. Here's my brief overview.
Horspool published some work on using (a library of) short oligonucleotides in DNA ligase reactions to build up to longer molecules:
1) Efficient assembly of very short oligonucleotides using T4 DNA ligase ( https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2994885/ )
2) Gene synthesis by assembly of short oligonucleotides ( https://central.bac-lac.gc.ca/.item?id=TC-BVAU-16681&op=pdf&app=Library&oclc_number=1032959739 )
Max Berry (member of this mailing list) cited that work in his Isothermal Amplification Array technique, which shows how to generate a library of all possible 8mers and then use DNA ligase reactions to build longer oligos. Using a standard DNA array, the reverse complement of each 8mer is synthesized along with a recognition site for a nicking endonuclease. By adding a polymerase, the nicking enzyme, and dNTPs to a particular spot on the array, oligos are polymerized, cleaved from the template by the nicking enzyme, and strand-displaced by the polymerase which then synthesizes more oligos. Thus each spot can generate a pool of the desired 8mer.
Isothermal amplification array proposal:
https://diyhpl.us/~bryan/papers2/bio/Berry-IA-Pilot-Project.pdf
Ligase can be used to join short oligos into a double-stranded form as long as there is sufficient base-pairing, especially if multiple oligos are added at once. Ligase cannot function in this fashion on oligos <6bp but may be amenable to directed evolution to improve this, as well as its general efficiency in this reaction. Each oligo is modified by polynucleotide kinase (PNK) to ‘activate’ it for use by the ligase. An algorithm must be used to plan the assembly reaction to avoid palindromic oligos which would preferentially self-pair and not contribute to the extension.
Separately, he also made a related technique that he calls Recombinase-Mediated Assembly, where he uses RecA-like recombinase UvsX which spontaneously forms a nucleofilament around ssDNA and then "scans" nearby dsDNA for a homologous sequence, whereupon it will disrupt the dsDNA double helix, insert the ssDNA, and allow other enzymes like polymerases or ligases to concatenate fragments together. This is similar to Gibson assembly but benefits from active base-pairing recognition between the ends of the fragments, as long as an ssDNA “tail” is available on one end of the fragment. The tail can be generated by using oligos unmodified by PNK to stabilize the ligation of multiple oligos on only one side of the fragment. Using these properties, it should be possible to do one-pot assembly of an arbitrary number of fragments, "generating dsDNA many kilobases in length, with minimal errors".
Recombinase-mediated assembly proposal:
https://diyhpl.us/~bryan/papers2/bio/Berry-RMA-Pilot-Project-1.pdf
He gave a talk outlining his two proposals here:
video: https://www.youtube.com/watch?v=is5G3LYdzms
transcript: https://diyhpl.us/wiki/transcripts/hgp-write/2017-05-09/recombinase-mediated-assembly-and-isothermal-amplification-arrays/
I also note that Ribbon Biolabs published another technique using a library of short oligos (6 to 22 nt each) and some dsDNA ("with at least one overhang") and DNA ligase ( https://patents.google.com/patent/US11352619B2/en ). In their method, they seem to do phosphoramidite synthesis of a library of short oligos on multiple 3192-well plates, optimizing the order of the oligos on the plates based on the number of times they appear in the final sequence (so as to minimize the number of times they have to switch between plates during assembly), and then they go about using a liquid handler or fluidic device to pick and choose different oligos to include as the next step in their ligase assembly reactions.
Bryan & Max
Over the past few years, there have been a number of interesting proposals to use ligation for making long oligos. Here's my brief overview.
Horspool published some work on using (a library of) short oligonucleotides in DNA ligase reactions to build up to longer molecules:
1) Efficient assembly of very short oligonucleotides using T4 DNA ligase ( https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2994885/ )
2) Gene synthesis by assembly of short oligonucleotides ( https://central.bac-lac.gc.ca/.item?id=TC-BVAU-16681&op=pdf&app=Library&oclc_number=1032959739 )
Max Berry (member of this mailing list) cited that work in his Isothermal Amplification Array technique, which shows how to generate a library of all possible 8mers and then use DNA ligase reactions to build longer oligos. Using a standard DNA array, the reverse complement of each 8mer is synthesized along with a recognition site for a nicking endonuclease. By adding a polymerase, the nicking enzyme, and dNTPs to a particular spot on the array, oligos are polymerized, cleaved from the template by the nicking enzyme, and strand-displaced by the polymerase which then synthesizes more oligos. Thus each spot can generate a pool of the desired 8mer.
Isothermal amplification array proposal:
https://diyhpl.us/~bryan/papers2/bio/Berry-IA-Pilot-Project.pdf
Ligase can be used to join short oligos into a double-stranded form as long as there is sufficient base-pairing, especially if multiple oligos are added at once. Ligase cannot function in this fashion on oligos <6bp but may be amenable to directed evolution to improve this, as well as its general efficiency in this reaction. Each oligo is modified by polynucleotide kinase (PNK) to ‘activate’ it for use by the ligase. An algorithm must be used to plan the assembly reaction to avoid palindromic oligos which would preferentially self-pair and not contribute to the extension.
Separately, he also made a related technique that he calls Recombinase-Mediated Assembly, where he uses RecA-like recombinase UvsX which spontaneously forms a nucleofilament around ssDNA and then "scans" nearby dsDNA for a homologous sequence, whereupon it will disrupt the dsDNA double helix, insert the ssDNA, and allow other enzymes like polymerases or ligases to concatenate fragments together. This is similar to Gibson assembly but benefits from active base-pairing recognition between the ends of the fragments, as long as an ssDNA “tail” is available on one end of the fragment. The tail can be generated by using oligos unmodified by PNK to stabilize the ligation of multiple oligos on only one side of the fragment. Using these properties, it should be possible to do one-pot assembly of an arbitrary number of fragments, "generating dsDNA many kilobases in length, with minimal errors".
Recombinase-mediated assembly proposal:
https://diyhpl.us/~bryan/papers2/bio/Berry-RMA-Pilot-Project-1.pdf
He gave a talk outlining his two proposals here:
video: https://www.youtube.com/watch?v=is5G3LYdzms
transcript: https://diyhpl.us/wiki/transcripts/hgp-write/2017-05-09/recombinase-mediated-assembly-and-isothermal-amplification-arrays/
I also note that Ribbon Biolabs published another technique using a library of short oligos (6 to 22 nt each) and some dsDNA ("with at least one overhang") and DNA ligase ( https://patents.google.com/patent/US11352619B2/en ). In their method, they seem to do phosphoramidite synthesis of a library of short oligos on multiple 3192-well plates, optimizing the order of the oligos on the plates based on the number of times they appear in the final sequence (so as to minimize the number of times they have to switch between plates during assembly), and then they go about using a liquid handler or fluidic device to pick and choose different oligos to include as the next step in their ligase assembly reactions.
Bryan & Max
- Bryan
https://twitter.com/kanzure
https://twitter.com/kanzure
Reply all
Reply to author
Forward
0 new messages