Thursday, June 7, 2012

DNA Binding Proteins

I've been tweaking my Ultimaker a lot lately so I haven't done many protein prints. I did do a couple recently and I think they came out pretty good:

Once again I've been making structures with DNA in them because it looks so cool.

Here's Topoisomerase:

Notice that DNA strand running up the middle. This model is definately a lot better in person.
Topoisomerase is an enzyme that untwists DNA so it doesn't get broken by being too twisted. Have you ever tried to untangle a rope? Ever wished that the rope would just pass through itself to remove knots, and unwind itself to remove twist? Well DNA is a magical rope that does all those things, so whenever our cells need to divide they can easily untangle that mess and sort it out. This particular model is from pdb structure 1A31 (thing). It's a human type I topoisomerase, meaning this molecule is responsible for relaxing twisting stresses on the DNA. What it does is it cuts one strand of the DNA double helix, allows it to twist around the uncut strand, then reanneals the cut strand to make sure the DNA stays intact.
  As you can imagine it's pretty crucial that topoisomerase reanneals the strands exactly how they were. For type II topoisomerases the task is even harder; type IIs cut BOTH STRANDS of the DNA double helix and allow another double helix to pass through the cut. These are your typical knot-dissolvers, and for them its even more crucial to make sure the double helix goes back to the way it was. A single cut in the double helix can be fixed by DNA ligase. A double cut is much harder, because the two ends can diffuse away and information will be lost.

And here's Catabolite Gene Activator Protein (CAP):

CAP is a transcription factor from E. Coli (1CGP) (thing) that functions to turn on the lac operon and other alternative nutrition pathways when glucose levels are low. What a transcription factor does is activate or inhibit RNA polymerase from transcribing certain genes into mRNA. Almost everything that a cell does is controlled by what genes are being expressed, aka, which regions of the genome are being transcribed. A single transcription factor molecule (or a dimer, as in the case of CAP), can result in the production of a single strand of mRNA. That mRNA in turn can produce many proteins, and each one of those proteins can have a huge number of individual interactions with other proteins or small molecules. So the action of each transcription factor has a huge outcome for the rest of the cell.
   CAP binds two molecules of cAMP, one in each monomer, and the binding increases CAP's affinity to DNA. When CAP binds DNA it bends the double helix almost 90 degrees, and turns the adjacent gene on by activating RNA transcription. In that sense it is a bit rare because most bacterial genes are on by default, and require the binding of a transcription factor to turn them off. The bend in DNA is thought to be required for activating transcription, possibly by opening the DNA up to RNA polymerase binding.

Oh and another thing I've been working on is making a protein model kit, similar to chemistry model kits. I'm still working on the right way for the pieces to interact with each other but here's what I have so far:

Tuesday, April 17, 2012

RNA stuff

In case you (all one of you!) are wondering why I've been missing it's because of this:

So besides making awesome protein structures I've also been making a bunch of boring stuff like belt tensioners and other parts to improve my ultimaker and bring it to a functioning state. Well now that I've got it sort of figured out, I thought I'd print a bunch of protein structures WITH SUPPORT! Because support is the future, and cutting protein structures up takes time and effort.

The first thing I made was actually something I also made with Heffy but then decided to remake with support because I wasn't happy about the glueing together. And that is the T7 bacteriophage RNA polymerase! RNA polymerase makes RNA, and as such it is essential to transmitting messages from the nucleus to the cytoplasm, where proteins get made. In this case of course the T7 polymerase works in the cytoplasm transcribing virus DNA, and anyway it would be in a bacterium (hence bacteriophage) which doesn't have a nucleus. In any case this polymerase is really small and thus highly amenable to crystallization with DNA and RNA in tact, essentially caught in the act of transcription. Other, bigger eukaryotic (and prokaryotic! Bacteria care about their RNAs too) polymerases were not crystallized with complete double strands of DNA, so it was not as interesting to me to make those structures when they would not be "complete". Not to mention they are huge, and thus much harder to cut up.

The silver one was of course made with my ultimaker, printed with support and it came out very hairy because the ultimaker doesn't retract. It didn't quite finish too, as the ultimaker lost USB connection near the end of the print, so the top "nub" is not present. In any case the ultimaker print is a lot more contiguous (not obviously two halves glued together, thus easier to look at) but because of the hairy issue the surface quality is not the best. Heffy did much better on that count.

And since I liked how that DNA looked in the polymerase I thought I'd print out tRNA, because I think tRNA is super cool and though it's not a protein, it's pretty close and its surface structure is a lot more interesting than many proteins. tRNA is of course the physical embodiment of the genetic code, with an anticodon on one end and the corresponding amino acid at the other (anticodon at the bottom of this picture, amino acid at the top right). I hope to print out a ribosome at some point but at this scale the ribosome would probably be the size of my whole hand in that picture, relative to that tRNA. I'd definately want to sort out that hair issue before I make something of that size.

Oh here are the links to thingiverse as promised: polymerase tRNA

Wednesday, March 21, 2012

Actin part two: Formin

Like everything in the cell, the nucleation and elongation of actin filaments is highly regulated. Here to perform part of this regulation is the Formin family of proteins. Formins are cool because they stay associated with the growing end of an actin filament, stepping along the elongating filament as each monomer is added. In this way they protect the growing filament from being capped (and its growth terminated) by the abundant capping protein. Formins are usually big long proteins with many different domains, so the domain I've chosen to represent is the Formin Homology 2 (FH2) domain, because it's the one responsible for keeping the Formin protein stepping along at the growing end of the actin filament.
Here's a single FH2 domain from the formin bni1p (pdb code 1y64). The N-terminus is on the top right and the C terminus at the bottom. Check out the lasso domain! That is how the FH2 domain dimerizes; that lasso loops around the post (knob at bottom left) of another monomer. Actin binds in the middle of the resulting doughnut shape. The thin linker part was printed along with the main portion of the protein, and the lasso was made separately. Then I glued the lasso on, but the linker part broke so that's why I'm holding it like that.

This is how the two FH2 monomers would be oriented in the dimer. These are printed with the lasso wrapped around the post, but I didn't print the linkers because I wanted to glue the monomers together with something flexible to emulate how they would move in real life. I tried gluing a nylon rope but the glue I used didn't seem to take.. I'll update when I come up with something better. For now, imagine the linkers there.
And here's a dimer bound to two actin monomers (almost like in the pdb structure!) Imagine this sitting at the top of a filament, the FH2 domains walking up along the filament when new actin monomers get added.

So that's it for now; I'm building a new extruder with a smaller nozzle, so hopefully soon I'll get that working and I will be getting even better print resolution. Or it will explode/melt. Time will tell.

Tuesday, March 13, 2012

Actin Crazy

So I work in a lab that studies actin so why not print out the actin structure so I can learn more about it?

Here is a nice overview of actin and what it does. Briefly, actin is polymerized into filaments for lots of different purposes inside the cell.
Here are two structures that show actin bound to Dnase I (ATN1 on pdb; left) and actin bound in a filament (2Y83 on pdb; right). Notice how the pointed-end (that's the bottom in the picture) is more compressed in the one on the right, that is supposed to fit in a filament. Whereas the one on the left has the pointed end sort of splayed out, because there's supposed to be a DNase I bound to the DNase I loop (biologists are creative aren't they). Haven't had a chance to print out more of the right structure so I can make a filament, but I will do that as soon as possible.

Also I need to work out a concrete set of scale. I sort of eyeballed these and it shows, the one on the right is noticeably larger.

Woo! First protein structure print. You can see I was messing with options so the rightmost one came out stringy and hole-y.

Check it out! Half an actin filament! I need a better way to hold the monomers together.. double sided tape doesn't look great. Eventually I'll put pegs in them but it would also be great to be able to tack them on wherever, just for play purposes.

Monday, March 12, 2012

Mission Statement

As a biochemist I love looking at protein structures. But what I love even more than looking at something is having it, physically, in my hands to play with. Sadly you can't put together a physical protein out of a bunch of elementary plastic parts like those chemical model kits (darn chemists get to have all the fun), and if you tried to make a protein out of chemical model parts you would be broke pretty quick (and you'd need to move into a hangar).

So the next best thing is 3d printing. There are websites that will 3d print protein models for you, and other websites that will print any old thing for you, but they tend to be ridiculously expensive and if I was going to do this for real (with the intent of learning something about proteins from the process) I would want to 3d print lots of different things, for example proteins in different conformations, maybe at different scales with different ligands, etc. I want to be able to play with the concept, not feel constrained by the price tag every time (and thus afraid to make a mistake).

So, simply, what I've done is built my own 3d printer for the purpose of printing these models.
This is Heffy. He can move his axes up to 4 mm/s, and has a nozzle size of 0.8 mm. Print quality is steadily increasing, but as of now I can achieve 0.4 mm layers with a minimum feature size of 1 mm. The cartesian robot part is based off this, with these electronics and a stepper-driven extruder similar to this.

So I'll post pictures of what I've made along with a short description and a link to the stl I used to make them.