Thus we cannot escape the depressing reality that T2 will get shorter and linewidth will get bigger as we increase the size of the protein studied. The reduced T2 is not only a problem for linewidth, but also causes loss of sensitivity as coherence decays during the defocusing and refocusing delays (1/(2J)) required for INEPT transfer in our 2D experiments. The only ray of hope comes in the form of a new technique called "TROSY" (transverse relaxation optimized spectroscopy), which takes advantage of the cancellation of dipole-dipole relaxation by CSA relaxation to get an effectively much longer T2 value; we will briefly discuss TROSY at the end of this chapter.

12.2.5 Sample Size

Another consideration is the amount of sample (in mg) required for NMR. Because it is the concentration (mM) of molecules that determines the signal strength in NMR, as the molecular size increases the desired concentration of around 1 mM corresponds to larger and larger amounts of protein (Table 12.1). For a small protein (e.g., RNase at 12.6 kD) the sample size is around 6 mg. Talk to a biochemist or molecular biologist if you think this is a small amount of pure protein! Now move to chymotrypsin (22.6 kD), which would require 11 mg of pure protein to be soluble and monomeric in 0.5 mL of water. For fatty acid synthase, with 20,000 amino acid residues and 21 polypeptide chains, we would need to dissolve 1.15 g of protein in our 0.5 mL sample volume, clearly an impossibility.

The mention of multimeric proteins brings up another issue: symmetry. As we already said, the size problem in NMR is due to two factors: the complexity problem (too many resonances to fit in a fixed range of chemical shifts without overlap) and the linewidth problem (decreasing T2 with increasing size). If we have a protein consisting of 10 identical subunits arranged in a symmetrical fashion, there will only be 1/10 as many unique positions within the molecule and the complexity problem is reduced by a factor of 10. But the linewidth problem is still there because the molecule tumbles slowly due to its large physical size. This is also true if you want to study a peptide or protein bound to a phospholipid micelle, even if the micelle is fully deuterated and therefore "invisible" to NMR: the linewidth is determined by the tumbling rate of the entire molecular assembly, in this case the peptide plus the micelle.

12.2.6 Uniform Labeling

There are still other tricks available to extend the size limit of protein NMR. Uniform labeling with 15N allows us to bring in another chemical-shift scale—the 15N chemical shift—and to the extent that different nitrogens in the molecule have a variation or spread ("dispersion") of chemical shifts, we can "pull apart" the overlap by introducing a third dimension in our experiments: 3D NMR. A 3D experiment has the same beneficial effect on overlap that we saw in going from 1D to 2D NMR, and we can now "cram" more resonances into our spectra without ambiguity. Labeling both 15N and 13 C allows us to transfer coherence across peptide bonds and into the amino acid side chains with nothing but one-bond INEPT transfers. Because INEPT transfer requires a delay of 1/(27), the T2 loss associated with the delay is greatly reduced if we rely only on the very large (30-150 Hz) one-bond couplings. This is the basis of the "triple resonance" (1H, 15N, and 13C) experiments. With all of these improvements we can extend the limit up to around 25-30 kD. Keep in mind that determining the structure of a 30 kD protein is no picnic even with 15N and 13C labeling. Research groups that specialize in this sort of thing have one subgroup focused on sample preparation, one subgroup running NMR experiments, one analyzing the NMR data, and one doing structure calculations. Even so, the whole process can take well over a year to complete.

12.2.7 Deuterated Proteins

One of the things that shortens T2 in larger molecules is the dipole-dipole interaction, and the biggest and most abundant dipole around is 1H. One way to reduce the dipole-dipole relaxation is to replace the 1H with 2H (i.e., with deuterium, D). The magnet strength (Y) of deuterium is about one seventh of that of proton, so the dipole-dipole relaxation is much less. Even partial deuteration (e.g., 50% randomly distributed) will give a significant improvement. In the extreme case of 100% deuteration we would have no 1H signals, but even then we can exchange back the "labile" NH protons with H2O and have at least one proton per residue. This not only radically reduces the complexity of the spectrum, even within the NH region because aromatic protons are removed, but it does so at the cost of a great reduction in information content of the NMR data: only NH-NH interactions are observed in the NOESY.

In 3D experiments where 13 C SQC of a Ca carbon is evolving (equivalent to t1 in a 2D HSQC) we normally "decouple" the attached protons by inserting a 1H 180° pulse in the center of the evolution period to reverse any /-coupling evolution. We could accomplish the same thing by turning on 1H decoupling (e.g., waltz-16) during the 13C evolution period. In deuterated proteins, the one-bond 2H-13C coupling, though only about one seventh (~20 Hz) of the 1JCH, can lead to significant broadening of the crosspeaks in the 13C dimension, reducing their signal-to-noise ratio. The situation is even worse for a P-CH2 carbon. The solution is to apply broadband 2H decoupling during the 13C evolution period: linewidths of Ca and Cp resonances in the 13C dimension are significantly reduced, bringing the peak heights up and out of the noise.

12.2.8 Why Bother With NMR?

With so many disadvantages to large-molecule NMR, you might wonder why we all don't trade in our magnets and spectrometers for area detectors and start doing X-ray crystallography. There are many unique advantages of NMR, the most important being the lack of crystal packing forces. NMR structures are obtained in the native environment of biomolecules without the artifacts of neighboring molecules packed together in a crystal array. In some cases, the NMR structure has been shown to be quite different from the crystal structure, in a way that provides more relevant insights into biological function. Another advantage of NMR is that the feasibility of a project is known fairly early: all you need to do is put it in an NMR tube and record a *H spectrum. You can see right away if the linewidth and the spread (dispersion) of chemical shifts is good enough to proceed with 2D experiments, isotopic labeling, and so on. Crystallography requires getting a crystal, a difficult and time-consuming process, before any data can be obtained. There are also, of course, "uncrystallizable" proteins that can never be studied by X-ray diffraction. NMR also is a much more flexible technique, allowing for simple modifications of the medium such as pH adjustment, temperature changes, and addition of small molecules. For example, an active site histidine can be titrated to determine its pKa without interference of any of the other histidine residues in the protein. To study the binding of proteins to other proteins or to DNA we can just add the other biomolecule to the solution; the crystallographer needs to start all over and cocrystallize the molecules, usually leading to a completely new and different problem to solve. Finally, although the size limitations are severe for NMR, more and more large proteins are being discovered that are covalent combinations of a number of small, specifically folded polypeptide segments ("domains") connected by short, flexible linkers. The linkers can be cut by mild and selective protease digestion, releasing the domains as biologically active small proteins amenable to solution structure determination by NMR. More recently new techniques, such as residual dipolar couplings (RDCs), are being applied to determine the relative orientations of these domains in the whole protein.

NMR and crystallography should be viewed as complementary rather than competing methods. Different kinds of structural information can be obtained, and information gleaned from one technique can aid in the process for the other. Still, NMR is the "younger sister" in the family and must meet a higher standard of proof to justify big spending and job security in corporate research and development.

Relaxation Audio Sounds Relaxation

Relaxation Audio Sounds Relaxation

This is an audio all about guiding you to relaxation. This is a Relaxation Audio Sounds with sounds called Relaxation.

Get My Free MP3 Audio

Post a comment