Life's Ratchet: How Molecular Machines Extract Order from Chaos (37 page)

Once the promoter is found, the polymerase binds more strongly to the DNA. Now the motion becomes unidirectional. The polymerase starts transcribing, but only after a few wrong starts, which create junk RNA. To transcribe, the DNA strands need to be opened and unwound. RNA polymerase achieves this by creating a local loop of open DNA. This is different from replication, during which whole strands of DNA are separated and copied. The polymerase opens the DNA locally and quickly recloses it. Transcription is accomplished by matching each DNA base pair with a matching RNA base pair. Suitable base pairs are floating around in the form of triphosphates—such as ATP (which provides an A) and the corresponding GTP, CTP, and UTP. Using these supercharged nucleotides has a great advantage: The energized nucleotides not only supply bases to be
inserted into the growing RNA strand, but also supply energy, which fuels the polymerase machine as it tracks along DNA.

Block and his coworkers, whom we briefly met when discussing kinesin motion, also measured the motion of RNA polymerase using their optical trap. They found that the polymerase moves in steps of one base pair at a time (0.34 nm). As the polymerase rides along the DNA, it drags behind itself a growing strand of transcribed RNA. Despite the drag force from the dangling RNA fragment, the polymerase machine is highly processive, transcribing thousands of base pairs without falling off the DNA. In one experiment, researchers applied 30 pN of force to the trailing RNA strand, but the machine was undeterred and continued on its merry way. This force is much bigger than would be needed to pull apart DNA and RNA (at the transcription site, they are temporarily bound through base-pairing). Therefore, RNA polymerase must serve as a strong clamp to hold RNA and DNA together.

MAKING PROTEINS

The third step in the DNA life cycle is the translation of the DNA information into an actual protein. This is achieved by the mother of all molecular machines: the
ribosome
. A human ribosome consists of about seventy-five proteins and multiple long and complicated RNA strands, all bound together into two major units, which come together to form the ribosome. Why RNA strands? So far, all the machines we have discussed were made of proteins. The reason RNA is so important in our cells is that it is versatile: RNA is an information carrier, like DNA, but because of RNA’s greater flexibility, it can assume complex three-dimensional shapes, which can catalyze reactions like proteins. For this reason, many people believe that RNA came first—before DNA and proteins. This makes the ribosome, an RNA-based machine, especially interesting. Every living being on earth has ribosomes, and they all contain RNA. Every human cell contains about 100 million ribosomes, and tiny bacteria, like
E. coli
, contain about 10,000 of them.

The translation of genetic information into proteins requires a translator. The translation is handled by small molecules, which are part RNA and part amino acid and are called transfer RNA (tRNA). They work sort
of like a dictionary: When you try to translate a word from English into German, you look up the English word in the left column and then find the German word in the right column. If the words are
flood
and
Überschwemmung
, we find that the two words have little in common. Without the dictionary, we would never guess that they mean the same thing (my wife teases me that German words are always three times longer than the English equivalent). tRNA works the same way: On one end of the tRNA molecule is the three-letter codon sequence, and on the other end is the corresponding amino acid. There is no way to know that AGU stands for serine, or GGC for glycine. We need this molecular dictionary.
Figure 7.15
shows the transcription and translation sequence of DNA.

FIGURE 7.15.
Transcription and translation: how DNA is translated into proteins. Right: The DNA message inside the cell nucleus is transcribed into messenger RNA (mRNA) by a machine called RNA polymerase. The mRNA leaves the nucleus. Left: The ribosome takes the mRNA and matches it up with a suitable transfer RNA (tRNA), such that the anticodon on the tRNA matches the codon on the mRNA. The amino acid on the other end of each tRNA is then attached to the growing polypeptide strand, which subsequently folds into a functional protein. The newly manufactured protein can be a structural protein like actin, an enzyme, or a molecular machine.

The dictionary is also coded in DNA. Our DNA contains information on how to make enzymes called tRNA synthetases, which, as the name suggests, synthesize tRNA. These enzymes have special pockets, such that
only the appropriate codon matches with the appropriate amino acid. The tRNA synthetases always make sure that there is an ample supply of all combinations of tRNAs, so the ribosome can do its work. Using the tRNAs as types, the ribosome works like a typesetting machine. It takes the incoming message (in the form of mRNA, produced by the RNA polymerase) and then moves along the mRNA step-by-step and matches up each codon on mRNA with an appropriate tRNA. Once it finds a match, it clips off the amino acid attached to the tRNA, and attaches it to a growing strand of amino acids—the protein in the making.

Note that even though the ribosome is more complex than the other molecular machines we have encountered, almost all of them have one thing in common: They move along a track. Kinesin moved along microtubules, myosin along actin filaments, helicase and RNA polymerase along DNA, and the ribosome moves along mRNA. In most of these examples, the track is not permanently affected by the molecular machine. But some molecular machines literally eat the track they are moving on.

A couple years back, I was hosting the well-known biochemist Gregory Goldberg as a colloquium speaker at Wayne State University. Greg has a wonderful, dry sense of humor. At the beginning of his talk, he postulated “Goldberg’s law of biophysics”: “A physicist may be converted to a biologist, but the reverse transformation is forbidden by nature.” He was being generous: It is just as difficult for a physicist to understand biology as it is for a biologist to understand physics. Trying to enter a new field is seriously hard work. We speak different languages even when we talk about the same things.

Gregory Goldberg is a pioneer in the study of a particular class of large molecules,
matrix metalloproteinases
, or MMPs. MMPs are enzymes. The substrates for Goldberg’s MMPs are the constituents of the extracellular matrix. One example is collagen, a protein that forms strong fibers that give our bodies shape. Without collagen, we would all live life as “The Blob,” in a slimy puddle on the floor. Our cells are trapped like spiders in a tangled web of collagen and other extracellular matrix materials. To move through this web, cells have to somehow cut their way through.
This is achieved by certain MMPs, which break down collagen. Without MMPs, dead cells could not be replaced, or the cartilage in your joints could not be restored. But MMPs have a dark side. Cancer cells co-opt MMPs to allow the cells to spread, causing metastasis. Other MMPs are involved in forming new blood vessels that feed cancerous tumors. MMPs are therefore a target of cancer research.

Goldberg has studied MMPs for almost twenty years, together with other colleagues around the world (a group that has been referred to as the “MMP Mafia”—a prominent member of this “mafia” is my colleague Rafael Fridman, who got me interested in measuring MMPs with AFM). In 2004, Goldberg and his colleagues at Washington University made a startling new discovery. They were trying to understand how a particular MMP, called MMP-1, breaks down collagen and diffuses. The rate of diffusion is given by the diffusion coefficient, but to measure the diffusion coefficient of single molecules is tricky. Molecules are invisible in an ordinary microscope. But Goldberg’s physics colleagues had just the right technique.

Fluorescence correlation spectroscopy (FCS) is used to measure diffusion in polymers, nanoparticles, and liquids in confinement. The technique is quite easy to understand: A laser is focused on the sample and causes special molecules, called fluorophores, to emit light at a wavelength that is distinct from the laser’s. When there is no fluorophore in the laser focus, everything will be dark. When a fluorophore wanders into the laser light by random diffusion, there will be a tiny flash of light. Because the fluorophores randomly move into and out of the laser focus, the measured light will fluctuate. Knowing the size of the laser spot, scientists can determine the diffusion coefficient by analyzing the noisy signal coming from the wandering fluorophores. By attaching a fluorophore to a larger molecule, such as MMP, they can thus determine MMP’s diffusion constant.

Much to their surprise, Goldberg and his colleagues at Washington University found that their data did not fit any simple diffusion model. Instead, the molecules appeared to move in a fixed direction as they continue along the collagen fiber. How could this be? Molecules are subject to random thermal motion, which has no preferred direction.

As we have seen, directional motion requires the degeneration of free energy into heat. But MMPs are enzymes—they do not hydrolyze ATP to
move around. How could they move unidirectionally without violating the second law of thermodynamics? It turns out that MMPs get their energy by eating the track they are moving on. This is a good strategy: MMP-1 is supposed to break down collagen. If the enzyme simply randomly diffused, it would break the collagen strand at random locations. Instead, MMP breaks down collagen as it moves along, systematically creating large gaps in the collagen through which the cell can move.

Each break in the collagen presents a high activation barrier, keeping MMP from moving in that direction. It therefore has no choice but to move away from the break. It then catalyzes another break, moves away from it, and keeps going. Goldberg and colleagues called this the burnt-bridge mechanism. Imagine driving on a very long bridge. As you drive along, you throw an explosive out the back of your car. The way back is now obstructed by a large hole in the bridge. You have no choice but to drive away from the holes you create. Keep doing that, and you will have no choice but to travel in a fixed direction. This mechanism avoids violating the second law, because the breakdown of collagen supplies energy, which is used to direct the motion of the track-eating MMP.

From the examples we have discussed so far, it may seem that almost all molecular machines move along tracks (with the exception of the ATP synthase, which rotates). But many other molecular machines in our cells hardly move. Pumps are one example. Our cells are filled with water, proteins, small molecules such as sugars, and lots of ions. Cells need to control the amount of small molecules and ions in their interior—otherwise, they could be in great trouble.

When I was a child in Germany, we had a large cherry tree reaching up to the second-floor balcony of our house. In the summer, you could pick ripe red cherries from the balcony. After a rain, the cherries were often split open. The reason was osmosis. Cherry skins are permeable to water, but they do not let sugar escape. The excess of sugar inside the cherry creates a nonequilibrium situation. The sugar molecules want to dilute themselves with more water to match the sugar concentration outside the cherry. When it rains, the sugar molecules get their chance. While
they cannot escape the cherry, the imbalance of sugar inside and outside the cherry produces a pressure that drives water inside the cherry. The cherry swells up and finally bursts open.

The same would happen to cells if they could not regulate the amount of sugar, protein, water, or ions in their interior. Cells have a rather impermeable cell membrane, but this membrane is pockmarked with a vast array of specialized pores. Many of the pores are passive: They let certain molecules through if the molecules happen to diffuse that way. But some are machines—active pumps that pump certain ions or molecules at the expense of ATP.

Usually, ions will move across a membrane until they reach equilibrium. Passive pores will let ions diffuse, but the ions will only diffuse until equilibrium is reached. To create a nonequilibrium concentration difference between the inside and the outside of the cells, active pumping is required. One such pump, the
sodium-potassium pump
, moves sodium out of the cell and potassium into the cell, maintaining the typically high-potassium, low-sodium environment found in most cells. These pumps are extremely important for the cell. Our cells typically expend one-third of their energy to run their sodium-potassium pumps. In nerve cells, the fraction of energy used for these pumps can be as high as two-thirds.