Cell division is triggered by a chain of chemical reactions that begins outside the cell, passes through the cell membrane, the cytoplasm, and the nuclear membrane, and ends in the nucleus. Each link in the chain provides a chance to control or halt the cell-division signal—which is just as well. When the controls fail, cell division runs out of control; a tumor develops.
Most carcinogens cause mutations, and most cancer-causing mutations arise in genes that encode signaling proteins. Fortunately for humans and other organisms, those same carcinogens arouse a cell’s cancer-fighting defenses. Of the weapons in a cell’s defensive armory, none is more important than the protein p53.
Known as a tumor suppressor, p53 stalls cell division and then repairs damaged DNA. If the damage is too great, it can also trigger the cell’s latent ability to destroy itself.
The stretch of DNA that encodes p53 is no more or less susceptible to carcinogenic attack than any other gene. But if the p53 gene takes a hit, a cell becomes defenseless. Cells drawn from half the cancers that afflict humans contain impotent mutants of p53.
A protein’s function comes largely from its structure. When a protein fails to work, something structural is often awry. But elucidating p53’s structure, mutated or not, has proven difficult.
The p53 molecule consists of four identical subunits, each 393 amino acids long. One central section of each subunit folds to form the core domain, which binds to DNA. Another central section folds to form the tetramization domain, which controls the opening and closing of the four subunits around the DNA. But the subunits’ two ends don’t fold at all. Those loose ends and the looseness of the assembled molecule frustrate crystallization.
Thirteen years ago Nikola Pavletich of Memorial Sloan-Kettering Cancer Center in New York City and his coworkers used enzymes to excise p53’s core domain. They crystallized it and applied x-ray crystallography to determine its structure. Four other teams excised the tetramization domain and determined its structure. Now, Alan Fersht of Cambridge University in England and his collaborators have succeeded in determining the conformation of full-length p53 in its working state: bound or unbound to DNA in solution. 1
Their work combines three different methods: nuclear magnetic resonance (NMR), small-angle x-ray scattering (SAXS), and electron microscopy (EM). In previous work, Fersht and his collaborators had found that some p53 mutations prevent the core domain from folding (see Physics Today November 2006, page 24). Knowing p53’s full-length structure, says Fersht, could help pharmacologists find small molecules that shore up mutant p53 and restore its function.
The floppiness of full-length p53 makes the protein’s structure hard to determine even with methods like NMR and SAXS that don’t require crystals. At any instant, each molecule could have a different conformation. Fersht circumvented the problem by creating a mutant that has just enough additional stability to avoid blurring the structure while retaining the ability to bind to DNA.
To determine p53’s structure, Fersht turned first to NMR. In general, NMR is unsuitable for proteins as large as p53. The all-important resonant spins relax too quickly to yield clear structures. Thanks to its floppiness, p53 behaves like a smaller molecule. Still, obtaining the NMR spectra wasn’t easy. The multitude of isotopic substitutions and other preparatory steps took 12 years to perfect and apply. The effort yielded structures whose core and tetramization domains matched the excised structures determined by crystallization.
To gain insight into how p53 works, Fersht sought to determine the structure of p53 with and without its DNA substrate. For that, Fersht turned to two methods that provide big-picture views: EM and SAXS.
Although EM doesn’t require crystalline samples, it does entail immobilizing the molecules. Each of the 100 000 or so molecules in the electron microscope’s aperture shows up as a fuzzy blob. Sorting and averaging the individual images yields the structure with a resolution of about 30 Å.
SAXS, by contrast, works in solution. The randomly oriented molecules scatter x rays into a radially symmetric pattern. Sophisticated computer algorithms compare the pattern with those produced by candidate structures and find the best match. The resolution is also about 30 Å.
The accompanying figure shows three different views of the structure of p53 bound to a short strand of DNA that includes one of p53’s usual targets. The grey envelope corresponds to the EM structure. Fitted inside are the higher-resolution NMR structures of the bound DNA (purple) and p53’s folded components: the subunit’s core domains (blue and green) and their tetramization domains (orange).
SAXS yields a consistent structure of the p53-DNA complex but not of unbound p53. SAXS shows that p53’s four subunits move apart to form an open structure when they’re not bound to DNA. By contrast, the EM structure of unbound p53 resembles that of bound p53. Fersht says it’s likely that the immobilization step needed for EM nudges the unbound protein into a closed structure. He expects that the closed structure is also present in solution but at far lower concentration than the open structure, whose wide empty hole is more suitable for snagging DNA.
Whether those and other structural insights will lead to a cancer treatment isn’t clear. However, Fersht’s multi-technique approach could prove useful in studying other proteins that have unfolded domains. As recently as 20 years ago, biologists believed that unfolded domains serve only to link the functionally more important folded domains. Now, it’s recognized that unfolded domains are crucial to some proteins. BRCA1 (breast cancer protein) and alpha-synuclein (Parkinson’s disease protein), for example, both feature significant unfolded stretches.
In p53’s case, the floppy ends bind to proteins that work with and regulate p53. Thanks to the floppiness, the ends change shape to bind to several different proteins, not just one. The architectural mantra that structural biologists like to quote may need a slight edit: Forms follow function.
