Native Mass Spectrometry and the Use of IM-MS to Probe Monoclonal Antibody Structure
Native MS enables retention of key structural attributes such as noncovalent interactions within the tertiary and quaternary structures of large protein complexes and can be usefully applied to determine mass and stoichiometric information . The basics of native MS involve using electrospray ionization (ESI) or more usually nanoelectrospray ionization (nESI) with the analyte dissolved in an aqueous volatile buffer, commonly ammonium acetate or ammonium bicarbonate. Most mAb samples are not presented in these buffer conditions, so desalting spin cartridges may be used to buffer exchange the samples prior to analysis. These can also serve to help remove unwanted salts such as sodium and potassium from the solution, which would otherwise form adducts with the sample during ionization, leading to peak broadening . The ionization process is made as soft as possible with the use of lower spray potentials, lower source temperature, and reduced or no desolvation gas flow and by throttling the vacuum in the first desolvation stages of the mass spectrometer [112, 113]. Under these experimental conditions, noncovalent interactions can be retained, as evidenced by stoichiometry data and the preservation of protein-protein or protein-ligand interactions. The charge states observed with native ESI are generally lower, and the charge distributions are commonly narrower in comparison with analytical ESI, meaning data presents with lower charge, z (and higher mass-to-charge ratio, m/z). This can be used to probe mAb-antigen interactions, to examine heterogeneity and glycan content in the product, and most powerfully to check for formulation or storage-dependent aggregation [56, 95, 111, 114].
Native MS approaches coupled to IM-MS are hybrid experiments that directly probe the structures of biological macromolecules, such as mAbs, and their complexes as well as providing mass and stoichiometry [20, 115, 116]. Ion mobility separates ions by pulsing them through a cell filled with an inert buffer gas in the presence of an electric field. The time it takes for an ion to traverse the cell is based on the charge, size, and shape of an ion and its interactions with the buffer gas at a given temperature, resulting in a value termed the rotationally averaged collision cross section (CCS). When coupled to a mass spectrometer, ions are now separated according to two parameters, the mobility (K) and m/z. It is therefore possible to distinguish coincident m/z species based on their oligomeric order and protein conformations. IM-MS for intact protein analysis started in the mid-1990s led by a few groups who used homemade instruments, for example, Jarrold and Clemmer et al. [117, 118] and also following the development of the first commercial traveling wave IM-MS (TWIMS) instrument by Bateman and coworkers in 2006 [119, 120].
Several types of ion mobility-mass spectrometers are available; however, antibody characterization to date has predominantly been reported using TWIMS [20, 116, 121, 122]. More recently use of drift time ion mobility-mass spectrometry (DT IM-MS) has also begun to show promise in this field [115, 123]. For DT IM-MS, CCS can be extrapolated directly from the measured arrival time distributions via the Mason-Schamp equation . When entering the drift tube, ions experience a force imposed by the electric field and collisions with neutral gas molecules. If the applied field intensity (E = V/L) (where L is the length of the drift cell) is sufficiently low and the gas number density (N) sufficiently high, ions will move with a constant velocity called a drift velocity (vd = L/td). This velocity is proportional to the electric field intensity, with the ion mobility (K) being defined as the proportionality constant between vd and E:
From this experimental measurement of K, it is possible to obtain the experimental DTCCS of an ion of mass M in the presence of a buffer gas of mass m using
For TWIMS, TWCCS can be obtained from the arrival time distributions of each ion indirectly using a calibration method [124, 125].
Application of IM-MS to mAbs has been used to solve a number of analytical challenges, from disulfide heterogeneity detection  to quantitative analysis of the folding status of the protein  as well as to determine the CCS of intact IgG [21, 115, 116, 127]. Currently IM-MS provides structural information at the global level whereby different conformational families can be resolved. For site-specific information, complementary techniques such as top-down MS and fast photochemical oxidation of proteins (FPOP) can be applied to obtain information for conformational changes at the regional and residue levels, respectively .
As for HDX-MS, information can also be obtained using IM-MS regarding the induced conformational changes upon the binding of the epitope and paratope of the antigen and intact mAb, respectively . mAbs are inherently flexible, existing in many conformations, in equilibrium, with only subtle differences between them. These subtle changes are distinguishable by IM-MS, whereas traditional techniques are only capable of resolving large shifts in structure . Pritchard et al. discussed the importance of knowing which different conformers are present when looking at the interaction of antigen proteins with antibodies. Different conformers may bind in significantly different ways, and hence performing an immunoassay without taking into account different conformational families and their relative abundances could potentially lead to an inaccurate conclusion upon biological activity .
As previously mentioned, the biological activity of mAbs can be significantly affected by PTMs, such as N-glycosylation [45, 46]. These PTMs were probed by Damen et al.  who exploited the three-stage IM cell of the SYNAPT HDMS instrument to interrogate the N-glycosylation profiles of different tras- tuzumab batches. Glycopeptides fragmented within the trap cell were then analyzed by TWIMS. With the transfer cell voltage set low, no further fragmentation was induced, providing MS/MS-style spectra to demonstrate the cleavage of glycosidic linkages. The authors then increased the trap cell voltage to form second-generation fragment ions due to peptide backbone cleavage, comparable to MS3. This type of analysis was quick and informative, provides intact mass, and could be expanded upon in the future to further characterize the molecular heterogeneity of mAbs. Recent work by Ruotolo et al. used collision-induced unfolding (CIU) followed by IM-MS to demonstrate how varying levels of glycosylation affect the stability and/or conformation and hence unfolding behavior of mAbs; no glycosylation led to unfolding at lower collision energies compared with fully glycosylated equivalents . Through use of collision energy ramps, unfolding was induced, and heat map plots were generated to give visual representation of the unfolding patterns. This method also enabled different IgG isoforms to be distinguished based on their differences in disulfide bond quantity and/or patterns of disulfide bonding.
In 2010, Bagal et al. used TWIMS to resolve structural isoforms of IgG2 mAbs that differed only in their disulfide bond heterogeneity . Analysis of two different IgG2 mAbs showed the presence of 2-3 conformers. A deglyco- sylated version of the IgG2 mAb also maintained the multiple gas-phase con- formers. These findings, alongside structural modeling, provided strong evidence toward the presence of IgG2 disulfide bond heterogeneity. The classical disulfide connectivity within IgG2 of four interchain linkages is referred to as IgG2-A; however, two additional isoforms are also known [130-133]. IgG2-B refers to the disulfide bond arrangement when one bond is formed between a
CH1 domain cysteine instead of a hinge region cysteine on one side of the molecule, whereas if both sides are bound in this way then the isoform present is IgG2-A/B. The work by Bagal et al demonstrated the ability of IM-MS to identify and resolve these different isoforms of IgG2. More recently an IgG2 isoform with similar disulfide structure but different biophysical and biochemical properties to IgG2-A was discovered [134, 135]. Reports now refer to the original as IgG2-A1 and the new isoform as IgG2-A2 . These isoforms, along with IgG2-B, have been studied with HDX-MS by Zhang et al. . Their results demonstrated that the IgG2-B isoform was a more compact global structure with regions of increased solvent protection just above the hinge region in comparison with the A1 isoform. This suggests that the two Fab arms in the B isoform are pulled closer together due to the nonclassical disulfide bond arrangement and that intramolecular contacts, in the form of Fab-Fab interactions, are present. These results were in line with thermal stability data also.
Typically IM experiments for mAb characterization are performed using the TWIMS setup; however, Pacholarz et al. have demonstrated how DT IM-MS can be used to explore the gas-phase dynamics of intact IgG1 and IgG4 mAbs
 . The authors observed an increase in both the IgG isoform CCS values with higher charge state; as expected, however a larger increase was observed for the IgG4 mAb. Due to the linking of the IgG4 LC being positioned further away from the center of mass than in the IgG1 isoform, it was speculated that this would allow for increased flexibility of the Fab arms, hence a larger CCS. This finding was contradictory to those observed for solution-based IgGs 2 and 4 and therefore raises interesting questions for the effects of desolvation upon flexible proteins, with MS holding the key.
The flexibility of IgG Fab arms has led to the development of bsAbs whereby two different IgG half molecules (one LC and one HC) recombine during a Fab-arm exchange (FAE) process to form a chimeric bsAb. These hybrid proteins are of particular interest as they have the potential to simultaneously bind two different antigens. The monitoring of the FAE process and of bsAb formation was performed using TWIMS for the first time by Debaene et al.
 . Two humanized IgG4 antibodies were analyzed separately and then combined with the addition of a mild reducing agent to facilitate bsAb formation for CCS determination. The SYNAPT G2 IM-MS instrument was capable of differentiating between the three different glycosylated mAbs despite molecular weights that differed by only ~0.5%. In addition native MS was used to monitor the formation of bsAb as a function of time, demonstrating how MS can be used for kinetic measurements as well. With the scope for being able to identify intramolecular differences between structural isoforms and for recognizing structural changes induced upon antigen binding, IM-MS could become increasingly popular within the field of mAb therapeutic development.
More recently, Pacholarz et al. have implemented variable temperature IM-MS (VT-IM-MS) to probe thermal stability of therapeutic mAbs and their fragments . Enhancing the stability of biologics is aimed to reduce aggregation and improve production consistency. Protein engineering is used to shift the mAb away from an aggregation-prone state by increasing the thermodynamic stability of the native fold, which may in turn alter conformational flexibility . VT-IM-MS has been used to study the unfolding of monomeric proteins and more recently for the unfolding and dissociation of large protein complexes [138-141]. Pacholarz et al. used VT-IM-MS to analyze mAbs and observe changes in the conformations of isolated proteins as a function of temperature from 300 to 550 K. The temperature at which the maximum structural collapse occurs correlates remarkably well with loss of quaternary structure and the solution-based melting temperature (1m). Diversity in the extent of collapse and subsequent unfolding are rationalized by differences in the hinge flexibility and strength of noncovalent interactions at the CH3 domain interface among IgG subclasses. VT-IM-MS provides insights into the structural thermodynamics of mAbs and was presented as a promising tool for thermal stability studies for proteins of therapeutic interest.