Protein Origami

So, imagine taking a long string, shaking it around, and having it spontaneously snap into this shape:

There’s an amusing article in the New York Times about a “crowdsourced” solution to what is called the “protein folding problem”. Every different protein has a unique three-dimensional shape, and the proper function of every different protein is dependent upon that specific shape. Proteins are like long flexible strings of beads, where each bead is a different amino acid. The strings of beads acquire the shape critical to their function by folding up into curlicues and clumps, held together by a large number of very weak bonds between the beads. Different proteins have different shapes because they have different arrangements of beads, with subtly different interactions. We know the linear sequence of amino acid beads in thousands of protein strings; the “protein folding problem” is using this two-dimensional, linear knowledge to predict exactly what three-dimensional shape the string will fold up into.

The high-powered way to solve this problem is to build bigger and bigger computers that can keep track of the one-to-one interactions of hundreds of amino acids. A lot of power is required, since the number of interactions is a factorial function of the number of amino acids, and a typical protein has hundreds of amino acids. The crowdsourced solution to this problem is to farm it out to thousands of bored computer gamers; by making it into a addictive video game, you get people to do the same thing (click here for the game, and a more detailed intro to protein folding. Don’t say you weren’t warned). The supercomputer, the gamer, and the living cell all do exactly the same thing: try out trillions of different arrangements of atoms, and see which ones are the most stable. The supercomputer thinks its deep thoughts, while the gamer clicks a mouse to drag atoms around on a screen trying to get a good score. The living cell makes its string of amino acids, and random thermal motion jiggles the string around until it finds ever more stable states.

Thermodynamics has been called “the science of desire,” a way of understanding what the universe wants. Basically, the universe’s overwhelming desire is to shed as much energy as possible. As anyone who has tried to keep a house and garden tidy, order does not usually emerge from chaos without a significant input of energy. Usually, the universe wants disorder, and making a complicated structure—whether a garden or a cell or an elaborately folded protein—requires the input of energy. The descent into chaos is spontaneous, ardently desired by the universe, since it represents that energy leaving the system.

However, there’s another way that energy can leave a system. Remember, the universe wants everything, proteins included, at an energetic minimum. Put two positively charged amino acids in a protein too close together, and the protein’s got too much energy stored in it, like a compressed spring. Separate two amino acids that want to stick together, and there’s too much energy stored in it, like a stretched out rubber band. These intramolecular forces are minimized in a properly folded protein.

So really, the universe’s desire can be met two different ways. Energy could leave the system by increasing chaos, or it could leave the system by relaxing intramolecular forces. If lots and lots of energy leaves a system by relaxing intramolecular forces, then that will satisfy the universe’s desires enough that it will tolerate an increase in order. This is why proteins can fold spontaneously, and why ordered structures such as crystals can form without the input of energy.

A beautiful case of spontaneous order-from-chaos was recently seen in a type of bacterial protein called a porin. These are proteins that sit in the outer membrane of the cell and act like a pore in one’s skin. They allow the passage of material into and out of the cell, and occasionally modify things as they move through. Porins are also absolutely lovely to look at. Imagine a single ribbon of over 200 beads spontaneously arranging themselves like so:

This structure is called a beta-barrel. The cell starts with an unstructured string of amino acids (1); intramolecular attraction will start to bring individual amino acids together to form a sort of a two-dimensional zig-zag (2). The zig-zag is essentially a two-dimensional sheet (3 and 4), which can be curled around to make a barrel (5).

It’s a bit of an oversimplification to say that this folds up on its own without any help. No energy is required, true, but there are few enzymes that are necessary to help it achieve this shape. Proteins that sit in the membrane are in an “oily” environment, and the amino acids on their surface must also be “oily.” However, these proteins are made in the inside of the cell, a watery environment—and oil and water don’t mix. So, in folding these proteins, a couple of problems must be overcome. First, an oily protein must be made stable in water. Second, it has to be transported to the appropriate membrane, and only then allowed to fold. There are specific enzymes to help with each step. However, the identity of these helping remained uncertain. There were lots of candidates. A group at Harvard led by Daniel Kahne came up with an ingenious solution to sorting the enzymatic wheat from the chaff.

Kahne’s group started by using a porin that allowed molecules into the cell and cut them in half as they moved through. Then, they fed this porin a molecule that emitted light when it was cut in half. So, if the porin was folded up properly—if it had received all the necessary help from all the enzymes along the way from the inside of the cell to the outer membrane—it could easily be detected by flashes of light as it worked. Finally, instead of working with intact cells, they worked with little blobs of membrane. The enzymes suspected of helping out with the porin folding process could then be added, one by one, to these blobs of membrane. If the right combination of enzymes was present, there was light as the porin cut up its work. If the right combination of enzymes was absent, the porin would never fold up properly, never cut up its target, and there would be no light.

Kahne’s experimental system: the porin (red can) is embedded in a bleb of membrane. It cuts a target molecule as it moves it into the bleb, and causes it to emit light. Unfolded porin protein, along with various helper enzymes, can be added to this system.

One essential enzyme actually had the job of preventing the porin from folding. This enzyme is called a “chaperone,” as it escorts the unfolded protein as it goes from the inner membrane to the outer membrane. Without this protein, the porin tried to fold up before it got to the membrane—so, no light.

Once the unfolded porin protein was delivered to the membrane blob (or, presumably, to the outer membrane of the cell), an additional five enzymes were required. These enzymes are permanently attached to the outer membrane, and without any one of them, Kahne’s group saw no light from the activity of the properly-folded porin. No other proteins originally suspected in the folding of the porin were required. Also, no energy was required—no chemical energy from ATP, no potential energy in the form of a voltage difference across the membrane, nothing. The process is truly spontaneous.

Presumably, this is what is going on in the cell: The chaperone proteins (pink hexagons) coat the unfolded protein, preventing oily amino acids from interacting with the water environment between the inner and outer membranes. The chaperones also deliver the unfolded protein to some helper enzymes (blue squares and rectangle), which allow the unfolded protein to spontaneously fold into a working porin (red drum).

So what do those five proteins do to help the porin fold? As the authors of this study write, “Together, [these proteins] perform a chemical transformation that we do not understand…” Presumably, some of them must receive the unfolded protein from the chaperone proteins, while others must insert the unfolded protein into the membrane, and perhaps feed it in slowly enough that it can fold bit by bit. The energy to do this probably comes from the formation of the intramolecular bonds that hold the beta barrel together. We don’t know how these enzymes work, but knowing their structure might give us some clues--perhaps we can get some gamers to work on them. In the meantime, it's wonderful to think about a jumbled string spontaneously becoming such a beautiful structure as a beta barrel.

Chrisine L. Hagan, Seokhee Kim, and Daniel Kahne (2010). Reconstitution of Outer Membrane Protein Assembly from Purified Components. Science 328 890-892.