This nonenzymatic macromolecule is a generic, protein with two covalently attached flanking arms (or "tentacles," if you will) comprised of custom, 30 nucleotide oligos. This reaction component is designed to bind to the DNA, each arm binding the region flanking the restriction site. The generic protein sits atop to the restriction site, providing steric inhibition.

Before the digest, this macromolecule is added to the reaction, and temperature is raised to 70 or so to partially denature the DNA. The restriction enzyme is then added. Since the restriction enzyme cannot access the site underneath the octopus molecule, it will not cut at that site, or it will cut at a reduced rate compared to other, unobstructed sites.

The idea here is to stop a restriction enzyme from cutting in an inappropriate location, such as in the middle of your gene.
-- Cuit_au_Four, Jun 10 2008

Octopus proteins...visualised, not in this talk but in the project. http://www.ted.com/...p/talks/view/id/147
please peruse the project, this is great effort. [4whom, Jun 10 2008]

this guy has his finger on the button... http://www.ted.com/talks/view/id/80
@ 16: 30 he talks of the data and who should be using it. [4whom, Jun 10 2008]

Restriction protection using PNAs http://pubs.acs.org.../abs/bi000675e.html
Similarish method using a DNA analogue in place of the protein/oligo construct. [MaxwellBuchanan, Jun 10 2008]

{wishing I were remotely qualified to comment on this}
-- pertinax, Jun 10 2008


"MaxwellBuchanan to the Science aisle please! We have a biochemistry spillage!"
-- wagster, Jun 10 2008


And here I am.

Some clarification for those who would like it. Restriction enzymes cut DNA a particular (usually very short) sequence, such as "GAATTC" - a bit like a pair of scissors that cuts printed text wherever it finds the word "the".

Often, you want to cut DNA with a restriction enzyme, but want to prevent its cutting at some specific points (for example, you don't want it to cut at the "GAATTC" that's in the middle of a particular gene, but you do want it to cut all the other GAATTC positions).

Cuit's idea is (a) partly denature the DNA (partly peel apart the two strands so that things can get in) (b) add a protein (which is just being used as a "lump of stuff") to which are glued two DNA "arms"; their sequence matches the sequence either side of the GAATTC that you *don't* want to cut. (c) Cool things down again so the DNA strands re-unite. The hope is that the "arms" will unite with the corresponding sequences in the target DNA, thereby gluing the protein in place and protecting that part of the DNA from the restriction enzyme.

The analogy would be a sort of metal strip that sticks to the phrase "then the boat sank", thereby protecting that particular "the" (but not others elsewhere) from the scissors.

So, basically quite a good idea. Some possible flaws:

(a) it's not trivial to couple the DNA 'arms' to a protein (though not difficult either). But it needn't be a protein; it could just be a stretch of 'uncuttable' DNA between the two arms

(b) The DNA is only partly denatured at the beginning. This is tricky, but essential if it is to easily zip back together afterwards. However, this "re- zipping" is very fast - probably faster than the binding of the "arms" to the DNA. So, the DNA will zip back together before the "arms" can get in and stick.

A better solution would be to use DNA- like molecules which recognize their matching sequence, but bind more strongly to the DNA strand than a "normal" DNA strand would do. There are some molecules like this, but I can't for the life of me remember what they are; I think they're PNAs (peptide nucleic acids; synthetic molecules with some properties of DNA and some of proteins; yet they are neither). PNAs (or PNA/DNA paired strands) aren't cut by restriction enzymes.
-- MaxwellBuchanan, Jun 10 2008


You rely on oligos to hold your molecule in place. I am not sure why you want or need a protein in the middle. Just make it one oligo. In fact, if this worked, you could do exactly what you propose with your own gene sequence: make a bunch of oligos complementary to your gene and throw them in.

If (you protest) your oligos will not be sticky enough to protect your gene, then your custom molecule would not stick either.

It occurs to me that one could denature the DNA, throw in the oligos complemetary to your gene, partly renature, then add some DNA crosslinkers to stick the oligo there. If you have a great excess of oligos probably some will be where you want them even if the whole thing has not renatured. You will want to use a reversible DNA crosslinker.

Or maybe not. Presumably you are trying to figure out what is next to your gene. For that it would not really matter of these gene itself was crosslinked to an oligo.

You could label your oligos, making it easier to recover the fragment of interest after cutting with enzyme.
-- bungston, Jun 10 2008


//Just make it one oligo. // No, not if you mean a single fully complementary oligo. It will just anneal to one strand, and the resulting double-stranded DNA will be cut by the enzyme. The net result will be that you'll cut one of the two strands, within your gene; may not be a problem, but then again it may. You could always use an oligo which was mis-matched at the restriction site (so, no double- stranded restriction target is created). But better to use a non- cleavable analogue, which could have the advantage of binding more strongly (see PNA in anno above).

I believe there a number of strategies for protecting chosen restriction sites, though I've never needed to use them.

[EDIT] Just google "restriction cleavage protection" for some examples. I've posted a link to the first one I hit, which does indeed use PNAs.
-- MaxwellBuchanan, Jun 10 2008


//Just google "restriction cleavage protection" for some examples.//
I bet there were couple of dodgy results in that search!
-- gnomethang, Jun 11 2008



random, halfbakery