Joachim Pietzsch
To discover how proteins fold into the correct structure, scientists have designed ingenious methods that can analyse their structures and folding patterns.
The folding of a protein takes place on a scale that remains beyond our comprehension (see The importance of protein folding). The conundrum is how proteins manage to fold into the correct structure, given that the number of possible conformations that a protein could adopt is astronomical, and how this process is carried out within a tiny fraction of a second. So, scientists have had to develop sophisticated tools that can not only uncover the final structure of the protein but also track the way in which this structure is formed. The two most important methods for determining the structure of proteins are X-ray crystallography and nuclear magnetic resonance spectroscopy (NMR).
X-ray crystallography
This method acts as an atomic microscope, using X-rays instead of visible light to determine the three-dimensional structure of proteins. To determine the conformation of a protein with X-rays, the protein must be in the form of a crystal with a strictly ordered structure. This is the most difficult challenge in X-ray structural analysis, and the slowest step of the whole procedure. Some proteins are relatively easy to crystallize; for example, myoglobin — the first protein to have its structure solved by X-ray analysis — forms crystals within a few days after the addition of a salt called ammonium sulphate. However, it can often take months or even years to crystallize proteins. And many proteins, particularly those that are anchored in cell membranes, still cannot be crystallized. There is no ideal way of crystallizing proteins, as it is more of an art than a precise science.
The crystallized protein is then irradiated with X-rays. X-rays are used because the wavelength of radiation used needs to be as small or smaller than the distance between the atoms in the crystal lattice of the protein. For example, two carbon atoms that are bound to one another are 0.154 millionths of a millimetre apart, so X-ray wavelengths of between 0.05 and 0.25 millionths of a millimetre are used — depending on what level of information is required. Some of the X-rays then pass through the gaps between the atoms in the body of the crystal without any change in direction — but other rays strike the electron shells of the protein’s atoms and are deflected from their original path. The rays that are deflected in the crystal meet on the way to the detector and are captured on X-ray film (alternatively the X-rays can be recorded electronically with a diffractometer). To determine the structure of the crystal in full, the crystal is rotated while it is being zapped by the X-rays.
To the untrained eye, the resulting X-ray film shows a meaningless pattern of several thousands of black spots — there can be as many as 25,000 different diffraction spots obtained from a small protein (Fig. 1). But to an expert, elaborate calculations and modifications can translate this pattern into the protein’s structure. Interpreting the pattern is based on the fact that the strength of the deflection of a ray depends on how many electrons there are in the electron shell of the atom that deflects it — for example, a carbon atom deflects about six times as strongly as a hydrogen atom. Also, the height of a wave from a deflected ray — its amplitude — is proportional to the number of electrons in the atom that caused the deflection.
The phases of the various diffracted waves when they meet depend on the arrangement of the atoms that cause the deflection in the crystal (Fig. 2), and therefore the intensity and arrangement of the spots can be used to calculate the electron density in the crystal. The three-dimensional electron-density distribution is obtained from a stack of many two-dimensional sections. The diagrammatic representation of the electron-density contours can be compared to the contour lines on a rambler’s map; they can be used to reconstruct the actual contours of the crystal landscape.
This process proved to be just as difficult to put into practice as it is to describe in principle. It took more than twenty years until the principles of X-ray crystallography, first postulated in the 1930s, were used to solve the structure of myoglobin in 1959.
Nuclear magnetic resonance spectroscopy
This process, which was first used to determine the three-dimensional structure of a protein in 1984, relies on the fact that some atomic nuclei, such as hydrogen, are intrinsically magnetic. In a magnetic field, these magnetic nuclei can adopt states of different energy. Applying radio-frequency radiation can induce the nuclei to flip between these energy states, which can be measured and depicted in the form of a spectrum (Fig. 3).
The NMR properties of a nucleus — such as the energy difference between the orientations and therefore the frequency at which that nucleus absorbs energy — depends on its chemical environment. Magnetic nuclei are affected by each other as well as the applied field, both through chemical bonds and over short distances through space. This can be exploited to assign resonance signals to particular nuclei in a complex structure, and derive constraints for the distances that seperate them.
These properties have made NMR a valuable tool in the determination of three-dimensional structures. One advantage of NMR spectroscopy is that it can measure the distances between certain atoms of a molecule irrespective of its spatial orientation, unlike X-ray crystallography, for which a molecule with a unique spatial orientation in a crystal is the essential prerequisite. An NMR study, therefore, avoids the need to crystallize proteins. Liquid solutions of test material are sufficient, so as well as proteins, many other macromolecules can be investigated — although the larger the proteins, the less precise the findings are. Until recently, the upper limit for molecules to be investigated by NMR was a molecular weight of about 20 kilodaltons (just below the average size of a protein — ~30–40 kilodaltons). But, ingenious improvements in the resolution of the procedure have now pushed this limit to about 100 kDa, which means that most proteins are, at least in principle, accessible to NMR analysis.
Another advantage is that NMR spectroscopy measures its samples on a time scale from a billionth of a second to a few seconds, which is similar to the time scales that are involved in protein folding. NMR can therefore be used to measure the movements of individual molecular groups or whole domains of proteins directly and dynamically. Although the X-rays that are used in crystallography are also deflected instantaneously, overall a crystallographic determination takes minutes or even hours, as it is not always possible to record a measurable signal immediately.
The NMR method has therefore become the preferred technique for the experimental observation of protein folding. To permit proteins to be observed at the right time, the NMR procedure has to be combined with ultra-fast preparation techniques, as folding would be complete before any measurements of interest could be recorded.
But, as NMR is an in vitro technique, then strictly speaking, it is not protein folding that is observed in the test-tube, but re-folding. The protein under investigation is first denatured with the aid of certain reagents. Re-folding from the countless denatured variants of the protein to the single natural form is initiated by diluting the denaturing reagent with appropriate buffer solutions, which neutralizes its action. A similar effect can be produced by a sudden change in temperature. Ideally, mixing or the temperature change in the protein solution should be complete before the first step of re-folding starts. The experimental measurement can then start a few milliseconds after the start of mixing. Combined with suitable methods of detection — it need not always be NMR — experimental researchers in protein folding are now able to use temperature jumps or ultra-rapid mixing to probe down to a threshold of a few millionths of a second.
These two techniques have provided fascinating glimpses into the world of protein folding. And improvements in these techniques, as well as other methods (such as confocal microscopy, which has revolutionized the ability to look at molecules in living cells), have led some scientists to predict that they will know about almost all protein folds by the end of the decade. Soon, we will have all the information that we need to understand how these processes go wrong in some diseases (see Treating protein folding diseases) and how they can be treated.
