Restriction enzymes and TdT?

[Nathan McCorkle]

an idea that was mentioned this past weekend to me at the
biohacktheplanet conference, was, essentially:

don't worry about terminators, if TdT extends n number... can't we
just find a restriction enzyme that is sequence-independent, but which
always cuts just past 1bp of ssDNA overhang (assumes you have a dsDNA
that had been extended by TdT on just one strand).


--
-Nathan

[Max Berry]

Not sure I understand the idea...first, that'd be a nuclease rather than a restriction enzyme. So they're saying that TdT, given e.g. a bunch of G nucleotides in solution, will add a poly-G to the strand, then you'd just have a nuclease that chews it back to a single G, and you repeat that process?

My main confusion is that TdT adds nucleotides in the 5'->3' direction, so I don't get how you can add additional basepairs without I them getting chewed back. If TdT was 3'->5' I could see having a polymerase synthesize the second strand to block the nuclease, but otherwise...

[Nathan McCorkle]

On Tuesday, August 29, 2017 at 8:52:20 PM UTC-7, Max Berry wrote:

Not sure I understand the idea...first, that'd be a nuclease rather than a restriction enzyme. So they're saying that TdT, given e.g. a bunch of G nucleotides in solution, will add a poly-G to the strand, then you'd just have a nuclease that chews it back to a single G, and you repeat that process?

Yep

 

My main confusion is that TdT adds nucleotides in the 5'->3' direction, so I don't get how you can add additional basepairs without I them getting chewed back. If TdT was 3'->5' I could see having a polymerase synthesize the second strand to block the nuclease, but otherwise...

Yeah, or could you flood in the complement nucleotide and ligase? There should be at least a bit more affinity for that to link the single nucleotide instead of ligating to the other (blunt) end, I think. 

[Nathan McCorkle]

On Wednesday, August 30, 2017 at 12:11:56 AM UTC-7, Nathan McCorkle wrote:

 

My main confusion is that TdT adds nucleotides in the 5'->3' direction, so I don't get how you can add additional basepairs without I them getting chewed back. If TdT was 3'->5' I could see having a polymerase synthesize the second strand to block the nuclease, but otherwise...

Yeah,

(filtering the enzyme at each step, preferably using some micro/nano fluidic) 

[Max Berry]

Ok, I could see the nuclease approach working with some protein engineering. Mung bean nuclease is used for blunting overhangs, and you could maaaybe do a screen to find a mutant that leaves the last base. Like the channel is blocked by some bulky residue so it can't move all the way down to a blunt end.

The ligase, you're gonna have a lot more trouble with. I'd love one that had such a minimal basepairing requirement, but hexamers seem to be the limit right now, and even then the efficiency is terrible, so getting down to monomers is a massive effort.