Basic help with Protein-DNA and Protein-Protein modeling and engineering

[Anonymous]

Hello,

I have absolutely no idea how to get started using Rosetta, and would greatly appreciate help. I really do mean absolutely no idea how to get started.

I want to use Rosetta to assist in engineering serine DNA recombinases. These are fairly small proteins that bind to, cleave, and recombine DNA in complex with both the DNA and three other recombinases in a symmetrical tetramer arrangement. They consist of a usually sequence-specific C-terminal DNA binding region that is modular and which doesn’t need to be dealt with, and a catalytic region that is responsible for both recognition of a ‘core’ DNA sequence motif and the cleavage and recombination of DNA in that core DNA sequence with other core DNA sequences bound by the other recombinases(1). This catalytic region is the region I (and everyone else) am interested in engineering. Ideally we’d like to be able to engineer recombinase variants capable of sequence-specific interaction with every DNA sequence. Doing so would make recombinases a nearly ideal genetic engineering tool (far better in many respects than the current toolset of ZFNs, TALENs, and CRISPR/Cas nucleases).

I’d like to do two things with Rosetta:

1) Most important: model and modify the small catalytic domain (<200 AA) of various recombinases in complex with various DNA sequences. This can be done modeling only one catalytic domain and a small amount of DNA (generally 10 to 20 base pairs).

There are some solved crystal structures of recombinase/DNA complexes to help with this. There are also primary sequences of known mutants of some catalytic regions that have specificity for certain non-native DNA sequences. It seems that these things would be useful, but I don’t know what to do with them in Rosetta.

2) Modify recombinase-recombinase interacting domains to ensure correct assembly and interaction of the recombinases.

This will assist in minimizing off-target effects. Basically I want to create steric or charged hindrance between wrongly interacting recombinases. A difficulty with this is that all four recombinases have to interact with each other during certain steps of the recombination process, but they fortunately don’t all have to interact in exactly the same way with each other. There are solved crystal structures of recombinase dimers and tetramers that can help in this sort of engineering. I’ll need to model protein-protein interactions using at least two to three catalytic domains at a time.

I’m ultimately looking to design recombinases that can precisely edit out the sometimes lethal, often debilitating polynucleotide DNA repeat expansions in diseases such as Huntington’s disease. They’d basically be turning DNA sequences like this:

ATGAAGGCCTTCGAGTCCCTCAAGTCCTTC
CAGCAGCAGCAGCAGCAGCAGCAGCAG
CAGCAGCAGCAGCAGCAGCAGCAGCAG
CAGCAGCAGCAGCAGCAGCAGCAGCAG
CAGCAGCAGCAGCAGCAGCAGCAGCAG
CAACAGCCGCCACCGCCGCCGCCGCCGCCG

Into this:

ATGAAGGCCTTCGAGTCCCTCAAGTCCTTC
CAGCAGCAGCAGCAGCAGCAGCAGCAG
CAACAGCCGCCACCGCCGCCGCCGCCGCCG

My experience with Rosetta consists of successfully building version 3.4 in BioLinux, reading some of the online manual, and playing some games on FoldIt. I can program a little, but haven’t done much outside of two introductory courses (Fortran77 and Java6). I’m not afraid to learn.

(1) – A paper with pictures of recombinases, if that helps: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3616721/

[Anonymous]

If you’re just starting out with Rosetta, I’d recommend taking a look at some of the tutorials that are out there. The Meiler lab has some good ones (http://www.meilerlab.org/index.php/jobs/resources click on Rosetta Tutorials), although nothing that’s protein-DNA specific.

The next thing I’d look for is the demos and protocol capture for the protocols you might be interested in. (These may be available as a separate download where you got Rosetta from.) Particularly, take a look at the protein_dna_interface_design demo, and protocol_capture/2010/RosettaScripts/RosettaDNA examples.

You haven’t mentioned any of the nuclease engineering papers which came out of the Rosetta community, so I’d suggest looking at those if you haven’t already. Make sure to also check out the supplemental information of these, which hopefully should give you an easily followable guide on how they ran Rosetta for their problem. I’m thinking here mostly of the papers by Summer Thyme & David Baker (http://www.ncbi.nlm.nih.gov/pubmed/?term=Summer+Thyme) and Justin Ashworth and David Baker (http://www.ncbi.nlm.nih.gov/pubmed/?term=Justin+Ashworth+David+Baker)

For your point 2), depending on how you set things up, you may be able to ignore the DNA component, at which point it’s a protein-protein interaction issue. There’s a number of demos and papers on that topic. If you’re looking for specificity among a set of proteins, you may want to add multistate design to the mix as well ( http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0020937 )

And of course, if you have any further questions, feel free to post a question in the forums here.

[Anonymous]

Thanks a bunch rmoretti, you’ve given me a lot of exciting things to look at and try.

[Author]

Posts