Folding Forum

About two years ago on the old forum I asked if during protein folding whether Knot Theory could be
used to explain why proteins fold they way they do and if it could be used to unfold them also. Since
then, someone has has written an article about DNA and the application of knot theory. See this
article,

http://www.sciencenews.org/articles/20071222/bob11.asp

Thus if a knot can be formed very quickly with DNA having two loose ends, then the same will apply
to proteins which also have loose ends. I've read articles that state that if the loose ends are held
together and one throws or agitates the string bundle while holding the two ends, the string, rope
will never become knotted. The same should apply to DNA/proteins. Now if a string with loose ends
is intermixed with one where the ends are held together, will the two become knoted together?

Moreover, since the protein can be represented as a chemical formula, one could say that chemical
formulas that have loose ends can become knotted whereas one that forms a circle won't. Can one
then look at the periodicy of the H and O and C letters in the formula and view all the paths of the H's in
the formula as one string, all the paths of the O's as another string, all the C's as another string and
determine if the H's can become knotted, the O's knotted and the C's knotted independently of each
other and perhaps dependently on each other (entangled) such that the H knotted string and the O knotted
string along with the C knotted string how they can influence the knot formations between them.
Moreover, one can probably view the chemical bonding between the H knots at certain points with
respect to the O knots and likewise the C knots has temporary knots (quasi knots) as the strings vibrate
in solution which may come into close chemical contact at certain points along their lengths. Hence
I see two variations as the actual physical knots due to the various H, o and C knotted strings that together make the
new chemical [reaction] and the non-physical knots that one can construct due to temporary electronic bonding alone.
Just like the H,O,C periodicy forming a H knotted string, O knotted string and C knotted string, one can
make a knotted string based on the electronic bonds (kind of like a virtual knotted string or the electronic
image of the physical one represented by H,O,C strings).

The next step down from this is at the atomic level vs. molecular level of knotted strings which current physicists
already use to define this space via string theory.

Yeah, string theory isn't really applicable to protein folding. Like you don't build a bridge with general relativity, but I don't think you were implying that really. I didn't read the article yet but keep in mind that DNA forms a very consistent secondary structure (the 3 forms of double helices) as well as some internal baspairing in single stranded DNA. Also large circualar DNA molecules (like a bacterial chromosome) do become supercoiled and have ends (like a piece of rope is supercoiled twine).

So, having not read the article, DNA forming knots in vivo wouldn't happen especially in eukariotic organisms (things that aren't bacteria and archea) where DNA is specially packed into chromatin, but in vitro... maybe.

I'll have to read the article, it might be interesting.

You may want to take a look at the references the article refers to also.

I realize alot of the extra stuff I said for chemical formulas and electronic bonds is just made-up, but
I was just day dreaming as to other possibilities that people may not of thought of by applying the
idea of knots from the article to these other concepts to see if there might be a connection from another
viewpoint even it is a far-out rambling....

If one reads the reference in the article for knotted proteins (oct 2006),

http://www.sciencenews.org/articles/200 ... thtrek.asp

This is what it said,

A protein is typically a long chain of amino acids, intricately folded into a compact package. Interestingly, although abundant and complex in polymers, knots are rare and simple in proteins. For the most part, knotted proteins contain trefoil knots (represented in diagrams with three crossings). Only three proteins have been found with knots that have four crossings (figure-8 knot).

Now, Peter Virnau, Leonid A. Mirny, and Mehran Kardar of the Massachusetts Institute of Technology have uncovered the most complicated knot yet discovered in a protein—one with five crossings. They describe their findings in the September PloS Computational Biology.

This protein (human ubiquitin hydrolase) contains a knot that crosses itself five times

It's better that those interested read the article in the above link for more detail since I would be cut-n-pasting
all of it if I keep at it.


Of course there are more references with this article on proteins.

Question:
If enzymes are the ones that seem to be the most likely to have knots (and complicated ones at that) to
degrade or un-degrade a protein marked for destruction, can drugs be made that have knotty properties
such that they degrade/encapsulate a disease (that is there cells) so that they can't interact with anything
to create more harm?

An educated guess to the last question:

The quotations of the articles citing knotted proteins mentioned that very few proteins are knotted that they have found. Enzymes are proteins most of the time (or RNA, or RNA-protein complexes, i.e. a ribozyme). So, just because those knotted proteins they have found are enzymes does not mean that knots are at all common in enzymes.

I'm not an expert in protein folding, but from what I have learned thus far in my biology studies I can say that proteins are most commonly made up of combinations of alpha-helices and beta-sheets (those are the most common contributers to secondary structure), there is no knot in these structures. Of course there is a lot more out there than just helices and sheets, there's zinc fingers, "insert-hydrophobic-residue-here" zippers (ex. leucine zippers), so who knows what is possible.

Anyway, back to your question. You mentioned these knotted proteins being targeted for degradation, and most proteins that have enzymatic activities are unstable. There has to be some turnover for the system to be able to return change quickly to maintain itself. There are actaully defined sequences that contribute to instability (PEST sequences), so a protein is is as stable as it needs to be, no more, no less.

On to the part about drug design. I did take a bioinformatics course last term and the book has a bit on drug design in the chapter about modeling of protein structure on computers. The first thing that needs to happen is to understand the pathology of the disease or condition, the biological nature. This helps a lot in being able to design a chemical that you might be able to synthesize, or to be lead to a therapeutic that the body may already produce (look up interferons for a cool example, they help in killing viruses) and make through recombinant technology (like how you make insulin).

(an aside here, there are lots of neat and promising things coming from improving on what is already present in people, i.e. finding the protein making some, tweaking it and using it as a drug. examples being interferons, tumor necrosis factor, human growth hormone, all of which have been made and altered to tweak therapeutic value while getting rid of side effects, having a wealth of structure information for this proteins can help make them even better and lead us to new breakthroughs).

So, you find out what is causing the problem. Is it a mutation, a virus, a bacterium etc. etc. You may or may not already have a starting compound that could be found by pure luck, or you could have some epiphany when studying the disease and come up with something. After you have that starting point, or maybe a few options, you start developing testing, most likely in mice and then to humans later, you study its effect, and its mechanisms. So, if you find the natural agent that works and it proves to be safe and you've studied it, and side effects are minimal, well you might be rich. If you just have an understanding of the disease or the organism that causes it you have to go through the whole approach of studying structure and mechanisms and design a chemical that might interfere with the mechanism. That might be something like how penicillin works (by blocking crosslinking of sugars in petidoglycan which makes up the bacterial cell wall, so no new bacteria can grow) or by blocking a receptor on a cell surface that the virus uses to get into the cell, or by giving the person what their body should produce but doesn't.

So drugs are like throwing a stick in the spokes of a bike of the pathogen, or oiling the wheels of the bike for yourself. A drug doesn't consume a pathogen just by causing instability in a few proteins, since it can make more proteins. Plus putting unstable proteins near other proteins does not make the others unstable.

The whole idea of this thread with knotted proteins, is that they have a novel structure that is worth looking into because we aren't familiar with it. As far as the research proving useful for drug design... You never know. But its just not a simple thing.


I went on so many tangents there.

Cheers,

Ok, I read the first article. It pretty interesting, who would have thought there would be knot functions.

If you've been following this thread topoisomerase II might be of interest to you. The enzyme binds to DNA where it must be untangled it makes a double strand break, allowing a strand to pass through and reanneals the broken strand afterwards. This helps reduce knots and supercoiling during replication. This requires energy, that is what the ATP in the diagram is for. Topoisomerase I just makes a single stand break and allows DNA to untwist if under stress, it doesn't require ATP, and it doesn't really have anything to do with knots. Topoisomerase II is used for decatenation after the replication of the bacterial chromosome or plasmid. It unlinks the two.