Chapter 1: Modern Mass Spectrometry and Advanced Fragmentation Methods

The increased interest in biomolecular mass spectrometry in recent years has largely been driven by technological evolutions. Many of these new developments will be described in detail in other chapters of this book; however, basic aspects of some of the key enabling technologies will be briefly reviewed here, including electrospray ionisation, common mass analysers, and ion mobility spectrometry. The utility of gas-phase fragmentation in structure elucidation will also be discussed, and limitations of traditional collision-based fragmentation in biomolecular analysis will be described, outlining the need for novel, more advanced fragmentation methods.

In 2019, it had been exactly 150 years since Dmitri Mendeleev compiled the first version of the periodic table of elements. He did this by arranging the then-known elements (around 60) by increasing atomic mass and noting the recurrence of certain chemical properties. Today, the number of identified elements – including synthetic ones – has roughly doubled, and the periodic table is perhaps the most recognisable achievement of all of chemistry to a general audience. For many of us, it was the first thing we saw the very first time we set foot in a chemistry classroom. Because of how universally important the periodic table is, any 18-year-old with even a limited interest in science is likely to know that every element in this table is associated with a number that indicates its atomic mass, and that this can be easily used to calculate the molecular mass of chemical compounds. They might even know that most elements occur in different variants, called isotopes, that have slightly different masses, and that the unit for atomic and molecular mass, the Dalton (Da), is one-twelfth of the mass of a carbon-12 atom.

If our hypothetical young person is interested enough to pursue a scientific study in higher education, they will likely learn that it is relatively fast and straightforward today to determine the molecular mass of a compound using a specialised instrument called a mass spectrometer. For most users, mass spectrometry (MS) is merely a tool used to confirm the nominal molecular mass of a compound. While the technique can be used in this way, this only scratches the surface of its capabilities. By high-resolution, high-accuracy mass measurement, it is often possible to determine a unique elemental composition for ions with molecular mass up to a few hundred Da. At sufficiently high resolution, exact isotopic variants can be distinguished, i.e., it can be determined to which nucleus a neutron was added. This is due to the different binding energies for different nuclei, which can be measured by mass spectrometry due to Einstein's famous equation, E=mc². These differences are usually of the order of one to a few mDa. Going to even higher resolution, it becomes possible to elucidate reactions by monitoring the movement of individual electrons. This might be surprising, given that the mass of an electron is only 0.00055 Da; however, measurement of ions of up to several hundred Da with this level of accuracy is routine today using high-performance instruments. At some point in the future, it might even become possible to measure chemical binding energies directly from accurate mass measurement using the same principle as for nuclear binding energies. As one electronvolt corresponds to a mass equivalent of ca. 1.1 nDa, this is beyond the capabilities of even the best instruments in use today.

Of course, while a significant amount of valuable information can be obtained from accurate measurement of the mass of an intact ion alone, the real strength of modern biomolecular MS lies in its ability to isolate an ion in the gas phase, break one or more of the chemical bonds present in the ion, and then measure the mass of the fragments. For both precursor and fragment masses, the above-mentioned benefits of accurate mass determination apply. A number of excellent introductory texts to mass spectrometry exist^1,2  and a familiarity with basic concepts is assumed in this book. Some key concepts such as commonly used ionisation methods and mass analysers in modern biomolecular MS, and figures of merit in MS will be briefly recapitulated here.

A mass spectrometer essentially consists of the following components: (1) an ion source, (2) a mass analyser, (3) a detector, and (4) a data acquisition system. Fundamentally, all mass spectrometers measure the mass-to-charge, or m/z value of gas-phase ions. Generally, this is because the interaction of the ions with an electromagnetic field accelerates them, while their inertia (i.e., mass) causes them to resist this acceleration. The magnitude of acceleration is proportional to charge, and inversely proportional to mass, and this is why, even though a variety of mass analysers have been developed relying on a range of physical principles, they all provide a measure of m/z. This of course assumes that the ions have a sufficiently large mean free path that their motion is not meaningfully affected by collisions with background gas molecules during mass analysis. For this reason, measurements are performed in a vacuum, and the vacuum system is usually either listed as a separate component, or understood to be part of the mass analyser.

Producing gas-phase ions is not trivial, and a range of methods has been developed over the years. Electron (impact) ionisation was developed in the early 20th century, first for solids,³  and later for gaseous⁴  samples. Unfortunately, this method is not appropriate for polar, high molecular mass analytes such as most biomolecules. Today, solid samples are more often analysed using matrix-assisted laser desorption/ionisation (MALDI),^5,6  secondary ion mass spectrometry (SIMS),⁷  or laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS).⁸  For biomolecular MS, samples are most commonly introduced in solution, and the most effective ionisation method in this case is (nano-) electrospray ionisation (ESI) (see Figure 1.1).^9,10  In this method, a potential difference is applied between a capillary containing the sample solution (at atmospheric pressure) and the inlet of the mass spectrometer. The resulting electrostatic force leads to deformation of the liquid surface to form a Taylor cone, which, when the attraction becomes sufficiently large to overcome the surface tension of the liquid, releases small droplets with an excess of charge. The exact mechanism for what happens next is still debated,^11–13  but this process ultimately results in formation of desolvated, multiply protonated (in positive ion mode) or deprotonated (negative mode) ions in the gas phase.

The two most commonly cited figures of merit in mass spectrometry are mass accuracy and resolution. Mass accuracy is simply the difference between the expected and measured m/z value. This is typically expressed as a fraction of the expected value rather than an absolute number, and as it tends to be a very small deviation using modern instrumentation, it is usually expressed as parts-per-million (ppm). It is important to emphasise at this point that this use of ‘ppm’ refers to accuracy of measured m/z values, and is unrelated to concentration or sensitivity. For example, the exact mass of singly protonated arginine (with all elements in their lightest isotopic variant) is 175.118952 Da. If our measurement is one ppm (or 0.0001%) too high, we will instead obtain a value of 175.119127, i.e., a deviation of less than 0.2 mDa, or 30% of the mass of an electron. This level of accuracy is typical for commercial Fourier transform ion cyclotron resonance (FTICR), orbitrap, and time-of-flight (TOF) instruments which are able to record data accurate to a few ppm (vide infra). The general formula is:

As the calculated mass is subtracted from the observed value, a positive ppm error is associated with overestimating the mass, and a negative value with underestimating it.

Resolution is also referred to as resolving power, and some authors interpret these two terms as having a subtly different meaning. In the context of mass spectrometry, these terms refer to peak width. Except in very specific single-ion applications, a signal in MS is always generated by many ions. Due to inherent instrumental factors including field inaccuracies and signal measurement errors, ions with identical m/z will be observed at slightly different m/z values, resulting in a signal in the spectrum possessing non-zero peak width. It should be noted that some commercial instruments actually do display zero-width bars (known as a ‘centroid spectrum’); however, in this case spectral processing occurs in the background to return a bar with an m/z value and height corresponding to the weighted average m/z value and intensity of the measured peak, respectively. Resolving power in MS is defined as the dimensionless ratio of the centroid m/z value of a peak, and the full width at half maximum of this same peak (Figure 1.2). Typically, this ratio will change as a function of m/z, in some cases in a very predictable way; for example, in FTICR systems the resolving power diminishes significantly as m/z increases. For this reason, it is good practice to report the m/z value where resolving power was measured or set. The primary benefit of high resolution is the ability to distinguish signals from ions with similar m/z in ‘crowded’ regions of a mass spectrum.

Finally, a less often mentioned characteristic is the acquisition rate of a mass spectrometer, i.e., how many spectra can be acquired per second. This varies considerably between the mass analysers typically used for modern biomolecular MS, from thousands of spectra per second for a time-of-flight instrument (although usually the raw unprocessed spectra are averaged to produce tens of spectra per second) to several seconds per scan for high-resolution FTICR measurements. It should be noted that the resolving power of FTICR and orbitrap systems reduces significantly as the scan speed is increased, unlike time-of-flight analysers in which the resolving power stays the same. When designing an experiment, it is important to take this acquisition rate into account, especially when mass analysis is coupled to online separation such as liquid chromatography. Using a mass spectrometer that is too slow in this case can result in not having enough spectra to properly characterise a chromatographic peak, potentially losing some of the benefit of the separation or even ‘missing’ very narrow peaks altogether.

In this section, we will briefly discuss the operating principles of some of the most prominent types of mass analyser in use today. Conceptually speaking, time-of-flight instruments are probably the simplest type of mass spectrometer in use today. Ions are accelerated by an electric potential difference, gaining a kinetic energy proportional to their charge. At the same kinetic energy (same charge), ions with a greater mass will reach a lower velocity. This can be expressed mathematically as:

where K is kinetic energy, V is acceleration potential, v is ion velocity, L is the length of the flight tube, and t is the time it takes the ion to traverse the flight tube. As V and L are constant, measurement of t yields the m/z value of the ion, and this is achieved by placing a detector at the end of the flight tube, which generates a signal when ions impact it. Limitations in resolution are due to the ions having opposing velocities within the acceleration field at the start of the experiment leading to the so-called time-of-flight turn-around-time aberration. Field imperfections can lead to varying kinetic energy in the flight path. These issues can be addressed; for example, a reflectron increases the flight time of ions and also compensates for the differences in ion kinetic energy spread. Of the three dominant types of mass analyser used in modern biomolecular MS (TOF, FTICR, orbitrap), TOF instruments have the lowest ultimate resolution when operating at very slow scan speeds (several seconds); however, as already pointed out, at more typical speeds of analysis, TOF systems are still able to operate at resolving powers greater than 100 000 with good mass accuracy. Therefore, benefits of this instrument type include an extremely high acquisition rate and nearly unlimited m/z range. TOF instruments with a range up to 100 000 m/z and beyond have been commercially available for some time.

In FTICR mass spectrometry, ions are stored in a Penning trap. A magnetic field (represented by the symbol B in Figure 1.3A) is applied along the Z axis of the trap (capitalised in this discussion so as to avoid confusion with the symbol for charge, z, in MS), and the motion of the ions is restricted along this axis by an electric field from trapping plates (Figure 1.3A).¹⁴  Any motion of an ion not parallel with this axis will by definition have a component that is perpendicular to the magnetic field, resulting in the ion experiencing a Lorentz force that will transform the original velocity component in the XY plane into an orbit. As the centripetal force to stabilise this orbit must equal the Lorentz force, the orbital motion of the ions is described by the following equation:

where v is the tangential velocity of the ion, r is the radius of ion motion, B is the magnetic field strength, and ω is the angular frequency of the ion. Therefore, rather than separating the ions in time or in space and having them hit a detector, it is necessary to measure the frequency of their cyclotron motion, preferably over many periods. This is done in practice by measuring the image current that the ions induce by approaching close to the walls of the ICR cell. The periodic signal detected this way is then Fourier transformed from the time domain to the frequency domain, and the resulting frequency spectrum is then easily converted to m/z. Importantly, the resonant frequency only depends on m/z, and not on the kinetic energy of the ion. In fact, increasing the kinetic energy, and therefore velocity of an ion in an ICR cell, leads to a proportional increase in r, and in other words only results in the ion being ‘excited’ to an orbit further away from the cell centre. This can be exploited in several ways for ion manipulation,^15–18  and is in fact needed to bring the orbiting ions close enough to the receiver electrodes for efficient image current detection; in other words, an ‘excite/detect’ workflow is used, much like in many other Fourier transform-based analytical techniques such as e.g., nuclear magnetic resonance. Due to the fact that the resonant frequency is independent of initial kinetic energy, extremely high accuracy (routinely below one ppm) and resolution (hundreds of thousands to millions) can be obtained in FTICR-MS. Higher resolution can typically be obtained simply by measuring the periodic image current for a longer time; however, in practice there are limits to how long a coherent signal can be maintained. Factors that ultimately cause the signal to decay include ‘space charge’ effects (this refers to the fact that the effective electric field that an ion experiences is affected by the fields caused by other nearby ions) and collisions with residual background gas. Further complications in practice are due to the trapping field, which leads to both a periodic back-and-forth motion along the Z axis, as well as precession of the centre of the cyclotron motion in the XY plane, known as the magnetron motion. One downside of FTICR is that the required magnetic field strengths are very high (12 tesla is fairly typical today) requiring the use of superconducting coils. These represent a significant expense, both to purchase and to continuously (not just during acquisitions!) keep at temperatures a few degrees above absolute zero, and there are also significant safety aspects that need to be taken into account when installing and operating such an instrument.

In the orbitrap mass analyser, developed by Makarov et al. in the early 2000s, only electrostatic fields are used to trap ions.^19–21  As a side note, there is currently only one vendor (Thermo, Waltham, MA, USA) of this type of instrument on the market, and as a result, the generic term ‘orbitrap’ and the brand name ‘Orbitrap’ are used interchangeably by many authors. In this analyser, a potential difference is applied between a central spindle-like electrode and a split outer barrel-like electrode, in such a way that the ions of interest will be attracted to the central electrode (Figure 1.3B). The ions, when injected properly, will therefore orbit around the central electrode (hence the name of the instrument) while moving back and forth between the ‘top’ and ‘bottom’ of the barrel. We will designate the symmetry axis of the orbitrap as the Z axis, so that the circular orbit of the ions is in the XY plane. As one might expect, the resulting ‘quadro-logarithmic’ electric potential is rather complex in Cartesian coordinates, but can be elegantly expressed in cylindrical coordinates as follows:

where r and Z are cylindrical coordinates (with Z=0 defined as the plane of symmetry of the field), k is field curvature, R_m is the characteristic radius, and C is a constant. While this equation might look fearsome, it is apparent upon inspection that the Z coordinate only appears once. Knowing that the potential energy of an ion in an electric field equals the product of the charge and the electric potential, the negative of the derivative of potential energy with respect to the Z coordinate, i.e., the force F_Z that the ion experiences in this direction, is given by:which is of course the equation describing a simple harmonic oscillator. Just like any harmonic oscillator, this is characterised by its resonant angular frequency ω, which is given byand so we see that m/z is proportional to ω⁻², so that it can be easily determined from image current measurement followed by Fourier transform, much as in FTICR. The orbitrap has seen tremendous developments over the past two decades and found significant applications in biomolecular mass spectrometry, offering mature data processing workflows^22–26  and almost FTICR-like performance, without the need for cryogenic superconducting magnets. It will be interesting to see in coming years if the still-existing performance gap between these two forms of Fourier transform MS will be bridged by further developments in orbitrap technology.

Magnetic sector instruments rely on transmitting a beam of ions through an orthogonal magnetic field after the ions have been accelerated by a potential difference. This again bends the ion trajectory, and, as in FTICR:

However, in a magnetic sector, only limited bending of the trajectory is performed, and the ions are not forced into an orbit. Substituting the expression for velocity derived in our discussion of TOF instruments, it is easy to see that

which simplifies to

As B and V are constant, the ions are separated in space (their motion is bent into partial orbits with different radii) depending on m/z. Magnetic sectors have been around since the first half of the 20th century and provide respectable mass resolution and accuracy; however, to obtain this level of performance, the use of physically large non-superconducting magnets is required to provide enough space for the trajectories of ions with different m/z to diverge sufficiently, resulting in relatively bulky instruments. More importantly, switching to a different m/z range requires changing the magnetic field strength B, and hysteresis therefore limits the scanning speed compared with a time-of-flight analyser. In practice, these instruments are mostly used today for targeted ultra-sensitive screening and quantitation of environmental pollutants such as polychlorinated biphenyls (PCBs) and dioxins. They are also employed in isotope ratio analysis, where low-mass ions in a narrow m/z range need to be separated.²⁷ 

The final type of common mass analyser that will be discussed here is the quadrupole. Here, an electric field is generated by superposition of a radiofrequency (rf) and direct current (dc) component on two pairs of opposing metal rods, with the two pairs offset by 90°. Ion motion in the resulting field is fairly complex and described by the Matthieu differential equation. As this behaviour is not easy to grasp intuitively, it will not be discussed in detail here. Depending on the rf and dc voltages used, only ions in a certain m/z window will follow a stable trajectory through the quadrupole, which therefore acts as a filter. Resolution and accuracy are low; therefore, simply placing a detector at the back of the quadrupole and registering whether a signal is measured as the window of stability is shifted along the m/z range of interest results in a low-performance, albeit robust mass spectrometer. Variants exist in which the ions are stored within the quadrupole by trapping plates (linear quadrupole trap), or where instead of four rods, two end cap electrodes and a ring electrode are used to create a 3D trap.²⁸  Given their low resolution, the utility of quadrupoles today is mostly as mass filters in so-called ‘hybrid’ instruments, in which, rather than a detector, a second mass analyser is placed after the quadrupole, or, more commonly, a fragmentation chamber followed by a second analyser.²⁹ 

Since the early 1990s there has been significant interest in the possibility to transfer intact, noncovalent biomolecular assemblies into the gas phase and analyse them by mass spectrometry. In 2004, the term ‘native mass spectrometry’ was coined for this approach,³⁰  and this method for MS analysis of biomolecules that are prior to ionisation in a native-like state has seen considerable evolution since its humble origins, to the point where even challenging analytes such as very large complexes or membrane proteins are tractable.^31–37  Good desolvation during electrospray ionisation without the use of organic solvents is critical to native MS, as is a relatively high tolerance to residual non-volatile salt, and for this reason nano-electrospray ionisation has been a key enabling technology for native MS.¹⁰ 

In addition to the stoichiometry of a noncovalent complex, it is possible to probe the physical size of an ion by measuring its gas-phase electrophoretic mobility. This is done through ion mobility spectrometry, historically also referred to as plasma electrophoresis (see Figure 1.4).³⁸  In this technique, an electric field is applied, often using a series of ring electrodes, causing an ion to migrate. In contrast to the MS analysers discussed in Section 1.4, this is not performed in vacuum, but in an inert gas, often He or N₂. Under these conditions, the ions will be accelerated by the electric field and slowed down by collisions with the background gas, reaching a macroscopic drift velocity equal to the electric field strength multiplied by mobility constant K. Under certain assumptions, including a low electric field strength, K is provided by:

where µ is the reduced mass of the ion–molecule collision complex, k is the Boltzmann constant, n is the number density of the background gas, and Ω is the collision cross-section of the ion. For large biomolecular ions, µ is essentially equal to the mass of the background gas, and arrival time is proportional to Ω/z. One commercialised variant of ion mobility spectrometry uses a series of alternating electrostatic potentials to generate a travelling voltage ‘wave’ through a stacked-ring ion guide,^39,40  complicating the relation between arrival time and cross-section.⁴¹  Coupling the exit of the ion mobility cell to the entrance of a mass analyser enables several interesting experiments, including simultaneous determination of m, z, and Ω. For native-like biomolecular assemblies, this can provide valuable structural information, as theoretical cross-section values can be calculated for each proposed geometry,^42–45  and it can then be determined whether these are consistent with experimental observations. Furthermore, changes in structure in response to solution parameters, ligand binding, and other factors can be studied in this way.^46–49 

Even at extremely high mass accuracy and resolution, only an elemental composition can be obtained from the mass spectrum of an intact ion. As described above, ion mobility spectrometry can provide some structural information and can be used to rule out certain geometries; however, structure elucidation by mass spectrometry generally requires gas-phase dissociation of an ion into fragments, as alluded to at the end of Section 1.4. In high-energy ionisation methods such as electron (impact) ionisation, unimolecular fragmentation of ions is commonly observed;^50,51  however, a range of activation methods has been developed over the years.^52–55  One of the simplest is collision-induced dissociation (CID), in which the kinetic energy of ions is increased and they are then made to collide with molecules of an inert gas, often helium, nitrogen, or argon. This process converts some of the kinetic energy into internal (vibrational) modes, ultimately leading to cleavage of covalent bonds. This method for ion activation has existed for more than half a century now,^56,57  has been implemented on a range of commercially available and homebuilt mass spectrometers, and can be considered a relatively mature technology. However, there are significant downsides to this fragmentation method, including a tendency for directed fragmentation at a limited number of preferred fragmentation sites, resulting in a relatively small number of fragments. Especially for large biomolecules with hundreds or thousands of covalent bonds, this can result in limited structural information. There are other limitations to CID; for example, certain types of isomer will show identical fragmentation behaviour and cannot be distinguished using this activation method. The redistribution of energy across vibrational modes that is typical for CID in the energy range found in most modern implementations also hampers investigation of noncovalent complexes, higher-order biomolecular folding, or labile modifications, as these are all typically lost early on in the fragmentation process. To overcome the limitations inherent to CID, considerable effort in recent years has gone into the development and application of more advanced fragmentation methods in biomolecular mass spectrometry. Rather than an inert gas, these are typically based on the interaction of precursor ions with electrons, photons, or solid surfaces. New dissociation pathways are opened up by these methods, enabling far more detailed structural characterisation by mass spectrometry.

Today, a wide range of mass spectrometry-based analysis techniques can be brought to bear on a bioanalytical problem. Given the variety of techniques, and the many ways they can be combined in hyphenated approaches, it is easy for the comparative newcomer to feel somewhat ‘lost’ in this area. Here, some of the basic concepts and technologies have been reviewed while keeping sight of the fundamental ion physics at play, so as to demystify the field to some extent. In subsequent chapters, the building blocks discussed here will be developed and combined in different ways in order to investigate the primary and higher-order structure of biomolecules by tandem MS.

This chapter has benefitted from insightful comments from Jeffery M. Brown.