What is NMR?
In short, NMR stands for Nuclear Magnetic Resonance spectroscopy. You may have come across it in A Level chemistry, using a 1D graph of thin, vertical lines to identify small (or large if you’re a chemist), organic molecules.
But what about using NMR to analyse large biological molecules over 1000× the size of the molecules you’ve seen in a year 13 chemistry exam? Whilst at first this may seem extremely complex and even unnecessary, NMR and its various applications have been a key driving force in the explosion of the field of structural biology over the past 70 years.
Before diving straight into the deep end, it’s worth considering the fundamental concepts that govern NMR and how this allows us to analyse a range of biological molecules from single amino acids to multi-protein complexes.
First a bit of physics…
At the heart of NMR is the concept of spin – the intrinsic angular momentum (i.e. the extent to which an object will continue to rotate in the absence of an applied force) of fundamental particles such as protons, neutrons and unpaired electrons.
Figure 1: The concept of the magnetic moment of a nucleus in an applied magnetic field.
Without going too much into the quantum mechanics behind fundamental particles, certain isotopes such as Hydrogen-1, Carbon-13 and Nitrogen-15 have spin-½ nuclei – a particular ratio of protons to neutrons that gives the nucleus a magnetic moment (i.e. the magnetic strength and orientation of an object) with two distinct energy levels when placed in a magnetic field.
Spin-½ magnetic moments can either align in a parallel (α) or an anti-parallel (β) orientation with respect to the direction of the applied magnetic field.
Figure 2: Schematic visualising the two energy levels of a spin-½ nucleus when placed in an external magnetic field – the higher energy, antiparallel orientation and the lower energy, parallel orientation. The energy difference between these two levels increases at higher magnetic field strengths.
Scaling it up to a whole population of atoms, most spins “point up” in the lower-energy, parallel orientation than “point down”, resulting in a net spin magnetisation along the direction of the magnetic field. This is useful because it allows us to imagine an ensemble of nuclear spins as a single vector rather than a complex network of spins all in different orientations at any given time.
Figure 3: The population of spins in both the parallel and antiparallel orientations can be summed up and conceptualised as a single magnetisation vector along the direction of the applied magnetic field.
So, we have established the nuclear and magnetic parts of NMR, but what about the resonance spectroscopy part?
In the simplest NMR experiment, a radiofrequency (RF) pulse is applied, whose frequency closely matches the Larmor frequency – that is, the frequency at which the nuclear spin “spins” about the magnetic field.
Figure 4: The application of a radiofrequency pulse stops the net magnetisation of the spins along the y axis (meaning that it is no longer at equilibrium).
This RF pulse is usually applied perpendicular to the direction of the external magnetic field and causes the spins to “resonate” – as a consequence, the net magnetisation (represented by the vector) is rotated 90° along the y-axis. Because the spins are no longer parallel to the powerful external magnetic field, they are no longer at equilibrium and tend to “relax” back to the equilibrium orientation.
Figure 5: After the application of a short RF pulse, the population of spins “relax” back to equilibrium, detected in the form of an electrical current along the y axis. This is known as the free induction decay (FID).
The relaxation of magnetisation back to equilibrium is detected in the form of an induced voltage in a detection coil, generating a signal known as the “free induction decay”. This signal is processed by a mathematics transformation known as the Fourier transform and gives rise to the common frequency spectrum that is interpreted and analysed. If we have a single sample of Hydrogen-1, for example, we would expect to see a peak at one frequency only, the resonant frequency of that nuclear spin.
Figure 6: The Fourier transform. Instead of leaving the complex free induction decay signal in the time domain, the Fourier transformation allows us instead to visualise our signals in the frequency domain.
Scaling it up…
So this may explain how we get a single peak in a 1D NMR spectrum, assuming we have a sample of a single isotope like the hydrogen atom… but what if we have something like an amino acid, or even a protein?
Well, NMR is exquisitely sensitive to local structure, meaning that the resonance frequency of individual spin-½ nuclei is highly dependent on the local electronic environment. This depends on bonding partners, bond lengths, bond strengths and so on. Hence, if a hydrogen atom is bonded to an oxygen atom, it will resonate at a different frequency to, let’s say, a hydrogen atom bonded to a less electronegative atom such as a carbon atom.
To understand why this is the case, we need to bring electrons into the picture.
Like protons and neutrons, electrons associated with atoms also react to an applied magnetic field. Electrons circulate about the direction of the magnetic field, creating a small, induced magnetic field at the nucleus. This can either oppose or reinforce the external magnetic field, resulting in a shift in the resonance frequency of that nuclear spin – what is commonly known as chemical shift (δ).
If the local magnetic field reinforces the external magnetic field (i.e. the circulation of electrons about a nucleus follows the same sense as the external magnetic field), then the nuclear spin resonates at a higher frequency. On the other hand, if the electrons oppose the external magnetic field, the nucleus is effectively “shielded” and thus resonates at a lower frequency.
Figure 7: Schematic demonstrating the induced magnetic field created at the nucleus by the circulation of electrons in an applied magnetic field.
The electron distribution of a nucleus depends strongly on what groups the atom is bonded to and therefore provides a convenient way of distinguishing between hydrogen atoms bonded to electronegative atoms like oxygen, and hydrogen atoms neighbouring less polar groups such as carbon.
In a 1D NMR spectrum, hydrogens belonging to methyl groups tend to be shielded due to the higher electron density at the nucleus and are therefore found “upfield” – to the far right of the spectrum. On the other hand, hydrogens belonging to hydroxyl or amide groups are deshielded due to the lower electron density at the nucleus (because oxygen and nitrogen nuclei are highly electronegative) and are found “downfield” – to the left of the spectrum.
One of the most interesting examples of chemical shift is the aromatic ring. If we consider the molecular orbitals of benzene, we find 6 π-electrons of the 6 sp2-hybridised carbon atoms delocalised above and below the plane of the ring. When placed in a magnetic field, the π electrons circulate to generate a strong “ring current” which opposes the applied magnetic field. Hydrogen nuclei at the edge of the ring therefore sense a stronger magnetic field and resonate at a higher frequency. This gives rise to the characteristically high chemical shift of aromatic protons (basically the NMR jargon for hydrogen atoms).
Figure 8: The ring current of an aromatic ring in an applied magnetic field strongly deshields the hydrogen nuclei at the periphery of the ring. Aromatic hydrogens are therefore found downfield on the NMR spectrum.
Some of you might be asking – why should I care about this? Let’s take a look at the scenario from a high-level point-of-view.
In biology, aromatic rings are found in the amino acids, tyrosine, tryptophan, histidine, and phenylalanine. These aromatic rings play an essential role in the primary, secondary and tertiary structure of proteins. Therefore, being able to distinguish different amino acids in such a sensitive manner therefore makes NMR a powerful tool when studying protein structure on both a primary sequence level and a tertiary structure level.
Going beyond chemical shifts
As you might be aware, chemical shift is not the only parameter in NMR. You may have come across peak splitting at A-level, a result of a parameter known as scalar coupling, or J-coupling.
Two adjacent spin-½ nuclei will experience an electron-coupled interaction which provides information on bond distances and angles in a molecule. This essentially allows us to determine which atom is bonded to which. This information is often contained in the complex splitting patterns of resonance lines in the NMR spectrum and becomes incredibly complex when the size of the molecule, and therefore the number of resonance lines, increases.
Figure 9: Schematic demonstrating the electron-coupled interaction between the two hydrogen nuclei (red dash) that gives rise to peak splitting.
For determining the molecular structure of individual molecules such as amino acids, peak splitting and J-coupling may indeed be useful. But generally, for molecules such as small proteins, peak splitting only tends to crowd the NMR spectrum, making it manually uninterpretable even to the trained spectroscopist. This is where the paths of the chemist and the biochemist diverge; whereas the chemist is content with a 1D spectrum to identify small molecules, the biochemist must go one step further to untangle the significant resonance overlap of small protein spectra.
Starting from the quantum mechanical concept of a spin, we have established how the specific magnetic properties of nuclei, namely spin-½ nuclei, allow us to probe the chemical and electronic environments of small molecules through the NMR parameters chemical shifts and J-coupling.
In the next part of this multi-part series on NMR, we will take a look at how biochemists overcome the complexity of 1D NMR spectra to obtain and assign spectra of proteins. As you will see, this will open up a wide range of experiments to investigate the various dynamic and structural properties of proteins and biomolecules, giving us vital insights into the function of proteins in the cell.
Author: Giacomo Casale, BSc Biochemistry with Research Abroad