Solving the Phase Problem in X-Ray Crystallography
This article explores the fundamental principles of X-ray crystallography and how physicists and structural biologists reconstruct three-dimensional macromolecular electron density maps. It details the mechanics of X-ray diffraction, defines the critical "phase problem" where vital wave phase information is lost during measurement, and explains the primary computational and experimental methods—Molecular Replacement, Isomorphous Replacement, and Anomalous Dispersion—used to recover this lost data and visualize atomic structures.
The Mechanics of Diffraction and Fourier Transforms
X-ray crystallography determines the atomic structure of macromolecules, such as proteins and nucleic acids, by firing a beam of X-rays at a crystallized sample. The electrons surrounding the atoms in the crystal diffract these X-rays, creating a specific geometric pattern of spots, known as reflections, on a detector.
Mathematically, this diffraction pattern is the Fourier transform of the molecule's electron density. Therefore, to reconstruct the 3D electron density map—which reveals the exact positions of the atoms—scientists must perform an inverse Fourier transform on the diffracted X-ray waves.
Defining the Phase Problem
Every diffracted X-ray wave is represented mathematically by a "structure factor," which consists of two critical components:
- Amplitude: The height of the wave.
- Phase: The relative position of the wave's peaks and troughs in space.
To compute the inverse Fourier transform and generate the electron density map, both the amplitude and the phase of each diffracted wave are required. However, X-ray detectors can only measure the intensity (brightness) of the diffracted spots. While the amplitude is easily calculated as the square root of the measured intensity, all information regarding the phase of the wave is completely lost during recording.
Because the phase carries the majority of the positional information about the atoms, this missing data—known in physics as the "phase problem"—prevents the direct calculation of the electron density map from the raw data alone.
Methods for Solving the Phase Problem
To reconstruct the electron density map, researchers must deduce the missing phases using indirect computational or experimental techniques. Three primary methods are used for macromolecules:
1. Molecular Replacement (MR)
Molecular Replacement is a computational method used when a structurally similar macromolecule has already been solved. Researchers use the known structure as a search model. Through rotational and translational algorithms, the model is positioned within the new crystal's repeating unit (unit cell) until its theoretically calculated diffraction pattern closely matches the newly measured intensities. Once aligned, the phases from the known model are borrowed and used as initial phase estimates for the unknown structure.
2. Multiple Isomorphous Replacement (MIR)
When no similar structure exists, phases must be found experimentally. In MIR, researchers soak the macromolecular crystal in a solution containing heavy metal atoms (such as mercury, platinum, or gold). These electron-dense atoms bind to specific sites on the protein and scatter X-rays strongly without altering the crystal's overall geometry.
Scientists measure the diffraction intensities of both the native crystal and the heavy-atom derivative crystal. By subtracting the native intensities from the derivative intensities, they can apply a mathematical tool called a Patterson function. The Patterson function uses only squared amplitudes (requiring no phases) to create a map of vectors between atoms. Because there are only a few heavy atoms, their positions are easily deduced from this vector map. Once the heavy atoms' locations are known, their phases can be calculated and used as a reference point to triangulate the phases of the entire macromolecule.
3. Anomalous Dispersion (SAD and MAD)
Normally, X-ray diffraction patterns are symmetric; reflections from opposite sides of a crystal have equal intensities, a principle known as Friedel’s Law. Single-wavelength (SAD) and Multi-wavelength Anomalous Dispersion (MAD) bypass the phase problem by intentionally breaking this law.
Researchers introduce a specific atom into the protein, most commonly by replacing sulfur-containing methionine with selenium-containing selenomethionine. They then tune the wavelength of the X-ray beam to match the exact energy absorption edge of the selenium atoms. At this specific wavelength, the selenium atoms absorb and re-emit the X-rays with a distinct phase shift. This phenomenon breaks Friedel's Law, causing symmetric reflections to have slightly different intensities. By measuring these tiny anomalous differences, computers can locate the selenium atoms via a Patterson map and directly calculate the initial phases for the macromolecule.
Map Reconstruction and Model Refinement
Once the initial phases are estimated using MR, MIR, or SAD/MAD, they are combined with the experimentally measured amplitudes. A computer then performs the inverse Fourier transform to generate an initial, often blurry, 3D electron density map.
Structural biologists fit a preliminary atomic model (the known amino acid or nucleotide sequence) into this electron density map. From this initial model, a new, more accurate set of phases is mathematically calculated. These updated model phases are combined with the original, accurate experimental amplitudes to generate a sharper electron density map. This iterative cycle, known as refinement, is repeated until the theoretical diffraction pattern of the atomic model converges with the experimental diffraction data, yielding the final, high-resolution 3D macromolecular structure.