Initial Protein Unfolding Events in Ubiquitin, Cytochrome c and Myoglobin Are Revealed with the Use of 213 nm UVPD Coupled to IM-MS
The initial stages of protein unfolding may reflect the stability of the entire fold and can also reveal which parts of a protein can be perturbed, without restructuring the rest. In this work, we couple UVPD with activated ion mobility mass spectrometry to measure how three model proteins start to unfold. Ubiquitin, cytochrome c and myoglobin ions produced via nESI from salty solutions are subjected to UV irradiation pre-mobility separation; experiments are conducted with a range of source conditions which alter the conformation of the precursor ion as shown by the drift time profiles. For all three proteins, the compact structures result in less fragmentation than more extended structures which emerge following progressive in-source activation. Cleavage sites are found to differ between conformational ensembles, for example, for the dominant charge state of cytochrome c [M + 7H]7+, cleavage at Phe10, Thr19 and Val20 was only observed in activating conditions whilst cleavage at Ala43 is dramatically enhanced. Mapping the photo-cleaved fragments onto crystallographic structures provides insight into the local structural changes that occur as protein unfolding progresses, which is coupled to global restructuring observed in the drift time profiles.
KeywordsUVPD IM-MS Ubiquitin Myoglobin Cytochrome c
Protein folding, the process by which an amino acid chain adopts a 3D structure is correlated to protein function and heavily regulated in the cell. Many pathways and intermediate structures form an accessible landscape towards a low-energy structure termed the native-state . Methods such as X-ray crystallography , nuclear magnetic resonance (NMR) , circular dichroism (CD)  and ion mobility-mass spectrometry (IM-MS) [5, 6] are often utilised to characterise protein conformation, and spectroscopic methods are employed to examine dynamics [7, 8].
IM-MS can readily report on conformational ensembles that a given protein can occupy at the same time [9, 10]. In ion mobility, ions progress through a buffer-gas-filled tube pulled by a weak electric field. Collisions with the inert gas molecules retard ions whilst the number and extent of collisions experienced depend on the ion’s size and shape, resulting in separation that is reflective of the ion’s size, shape and charge. Due to the electric field and collisions, ions ‘tumble’ and rotate as they drift through the tube, and ion mobility drift times can be converted into (or calibrated to obtain) a rotationally averaged collision cross-section, effectively a global structure measurement [11, 12].
The vast majority of studies of the structure of proteins in the gas phase utilise ESI with some modifications from methods used for rapid analysis of mass alone . Nano-electrospray (nESI) tends to be used more routinely for the analysis of biomolecules with the benefit of low sample consumption and lower spraying voltages, representing an even softer ionisation approach aiding the conservation of native-like conformations [14, 15, 16]. In order to retain a protein’s native structure in mass spectrometry (native MS), solution conditions have to be used that mimic the native environment. Aqueous salts with high volatility such as ammonium acetate can be used to prevent disruption to the non-covalent interactions present in solution. They act to stabilise the protein as it transfers from solution to the gas phase, whilst not being retained excessively such that a series of ions corresponding to different net charges is resolved.
Of course, the question presents itself whether the observed species in vacuo correspond to structures present in solution, and if they do, how far these similarities go. A number of mass spectrometry experiments have demonstrated that it is possible to at least partially retain solution structure and topology [17, 18, 19, 20, 21, 22, 23, 24, 25]. Careful selection of ionisation parameters has allowed analysis of macromolecular complexes keeping non-covalent interactions such as protein-ligand or protein-cofactor interactions , complete ribosomes  and even viral capsids  intact. Following electrospray ionisation, the mass spectrum of a given protein reflects the conformational spread that is present in the electrospray solution, prior to transfer to the gas phase. Typically, higher charge states and a wider charge state distribution are seen for unfolded proteins as the solvent-accessible surface area increases [28, 29, 30, 31, 32, 33].
Once in the gas phase, a plethora of activation techniques can be applied to biomolecules to obtain information on structure, sequence and potential interactions that may occur [34, 35, 36, 37]. The most widely used is collision-induced activation (CID), which has been used for both ‘bottom-up’ experiments where proteolytic fragments are analysed and for ‘top-down’ analysis of the intact protein . The CID process is slow and involves randomisation of energy throughout the molecule, such that is it difficult to obtain topological information from the fragment spectra, even when the fragments are subunits of a given protein complex. In contrast, topological information can be obtained using surface-induced dissociation (SID). This imparts energy to the molecule in a single, fast collision with a surface yielding subunits from complexes. The SID process can be optimised such that dissociation occurs prior to unfolding yielding far better insights to the arrangement of subunit in a given protein complex [39, 40]. Unlike CID, electron capture dissociation (ECD) is a non-ergodic process wherein capture and cleavage occur prior to statistical rearrangement of energy and can be used to probe structure . Unfortunately, ECD is less efficient for low-charge-state ions which are more reflective of the native fold [41, 42]. Electron-transfer dissociation (ETD) is also used for protein structure analysis although some of the caveats for ECD also apply . The efficiency of both ECD and ETD can be increased with the addition of CID however this will perturb the structure in vacuo.
An alternative approach couples laser activation with mass spectrometry termed photodissociation, whereby photons target specific regions within the protein and induce fragmentation [43, 44, 45, 46, 47]. Ultraviolet photodissociation (UVPD) utilises high-energy photons causing fragmentation in a single process . The nature of the cleavage and the ensuing photofragment will depend on the wavelength used. For example, at 266 nm, efficient UVPD requires the presence of an aromatic amino acid such as tryptophan, tyrosine or phenylalanine which acts as a chromophore to mediate backbone cleavage [49, 50]. At ‘deeper’ UV wavelengths, photons are also absorbed by the protein backbone . Brodbelt and co-workers have contributed a large body of work which in the main utilises 193 nm UVPD and have applied this to bottom-up as well as top-down analysis, including phosphate location and investigation of protein topology [51, 52, 53]. Stereo-chemical information on cis/trans isomerisation of proline residues in polypeptides has been found by Reilly and co-workers at 157 nm and von Helden and co-workers working at 193 nm [54, 55]. 213 nm UVPD has been used by two independent studies to fragment deuterium-labelled peptides both of which show successful retention of selective sites of deuterium labelling throughout the protein backbone, providing an alternative activation method to electron-mediated approaches [56, 57]. Girod et al. have recently shown that 213 nm UVPD of proline-containing proteins is able to provide sequence information including unusual fragment ions a+2 and b+2, previously seen with 157 nm UVPD, and attributed to prolines [54, 58].
The ability of UVPD at 193 nm to assess the conformation of proteins has been investigated previously by Cammarata and Brodbelt for holo- and apo-horse heart myoglobin on native-like charge states (8+, 9+ and 10+), showing that the fragmentation yield correlates with the rigidity of structural elements . Warnke et al. have incorporated ion mobility separation and selection with 193 nm UVPD, demonstrating that for ubiquitin 7+, irradiation of specific portions of the arrival time distribution produces both similar and distinctive fragmentation patterns that could indicate interconversion of species . Ly and Julian utilised 266 nm UVPD to cleave a carbon-iodide bond incorporated into Tyr59 of ubiquitin and subsequent radical-directed dissociation cleavage sites provided distance constraints for molecular modelling conformational searches, yielding structures that were distinctly different to the crystal structure and in agreement with data from ECD [25, 61]. We have also previously observed differences in the fragmentation pattern of two different, distinct conformational families of 5+ melittin as probed by 266 nm UVPD . Overall, there is increasing evidence that ultraviolet photodissociation, especially in combination with IM-MS, is a promising avenue for protein-structural characterisation.
In this study, we report how UVPD can be used to investigate changes in protein structure that occur during the transfer of proteins into the gas phase, by coupling a 213 nm laser to an ion mobility mass spectrometer. In turn, this shows how this approach could be used to provide topological information on the non-covalent interactions that stabilise protein fold. We use three well-studied proteins ubiquitin, cytochrome c and myoglobin and for each ion mobility data is used to determine ionisation conditions which preserve a compact native-like conformation termed ‘gentle’ or ‘soft’ to more elongated states termed ‘activating’ or ‘harsh’. As we alter the conditions, the appearance of distinct photofragments is monitored by mass spectrometry and ion mobility.
Methanol (> 99.9% purity) was purchased from Sigma-Aldrich (UK). Ammonium acetate was supplied by Fisher Scientific (Loughborough, UK). Ultrapure water was obtained from a Milli-Q Advantage A10 ultrapure water filtration system (Merck Millipore, Darmstadt, Germany).
Bovine ubiquitin, equine myoglobin and bovine cytochrome c were purchased from Sigma-Aldrich (UK) as lyophilized powders with purities of ≥ 98, 90 and 95% respectively. Proteins were dissolved in 200 mM ammonium acetate. Myoglobin in ammonium acetate was desalted twice using Micro Bio-Spin 6 chromatography columns (Bio-Rad Laboratories, Hercules, CA, US).
All samples were diluted to a final protein concentration of 10 μM.
Instrumental Setup for UVPD-IM
Modest modifications have been made to a Waters Synapt G2-S, in addition to those described previously to facilitate better injection of a laser beam in the trap cell region [56, 63]. Essentially, a CaF2 window was integrated into the source block and a hole was drilled into the stepwave cap allowing the laser beam to enter into the upper part of the ions’ pathway to the time-of-flight region where the beam exits via the mirror inserted in the push plate. UVPD is carried out pre-ion mobility separation in the trap cell region of the instrument. The general trapping procedure has been described previously  but in brief: a DC potential gate is applied to the trap cell therefore accumulating ions prior to the mobility cell; following this, the t-wave amplitude is set to zero volts and the ion influx is stopped by imposing a stopping DC potential on the trap cell entrance electrode. Subsequently, a mechanical shutter opens, allowing the laser beam to interact with the confined ions. The laser beam (CryLaS, Germany) operates at 1 kHz and delivers 213 nm photons with an average pulse energy of 2 to 3 μJ. After photoactivation, the shutter closes and the ions are pulsed into the ion mobility cell by re-enabling the t-waves in the trap cell and synchronising the fall of the DC potential gate with the IMS cycle. Ion mobility separation occurs and ions proceed to the time-of-flight detector allowing drift time measurement and mass analysis of both precursor and photoproduct ions.
All samples were analysed in positive ion mode using a nanoESI source. Borosilicate glass capillaries (World Precision Instruments, Stevenage, UK) were pulled in house on a Flaming/Brown P-1000 micropipette puller (Sutter Instrument Company, Novato, CA, US). Capillaries were filled with sample using a 10-μl Hamilton syringe (Hamilton Company, Reno, NV, US) and a platinum wire was inserted into the solution to allow application of positive voltage.
Spraying voltages were kept as low as possible to achieve gentle ionisation with a typical capillary voltage of around 1.1 kV, source temperature of 40 °C, and a trap bias settings of 30–35 V to reduce protein activation.
For the UVPD-IM experiments performed on the Synapt G2-S, the species of interest were m/z selected in the quadrupole, accumulated in the trap cell for 2 s in order to obtain an ion count of ~ 2e3 and ions were then photoactivated for 2 s before m/z and drift-time analysis. This procedure was repeated for each protein at different cone voltages ranging from 10 V up to 150 V. All UVPD experiments were done in triplicate over at least 2 separate days and results averaged between repeats. Normalised fragmentation yield per cleavage site (residue) was calculated according to this relationship:
aIM-MS experiments were performed on a Synapt G2-Si. The drift time of a mass-selected precursor ion was recorded whilst either the cone voltage or trap collision voltage was ramped in increments of 2 V from 10 to 150 V and 4 to 150 V respectively. Data analysis was done using MassLynx v4.1 (Waters Corporation, USA), ORIGAMI , OriginPro 9.1 (OriginLab Corporation, USA) and Microsoft Excel 2010 (Microsoft, USA).
Results and Discussion
Bovine ubiquitin was sprayed from a 200 mM ammonium acetate solution and observed as ions [M + 5H]5+ and [M + 6H]6+ (SI Figure 1). Upon altering the cone voltage from 10 to 120 V, no major changes were observed in the distribution of charge states in the mass spectrum.
Whilst fragmentation is already visible by the eye in the spectrum under ‘soft’ conditions, it increases dramatically with an increase in CV. The most intense fragments in both spectra result from cleavage at residues Glu18, Ile36 and Pro37, which are all on the N-terminal side of a proline residue. The effect of proline enhancing cleavage  has been commonly observed in MS/MS spectra of peptides and proteins . No fragmentation was observed in the N-terminal beta strand or in the C-terminal part of the protein after residue Lys48, as well as only few cleavages at the C-terminal end of the alpha helix (see Fig. 2c). Comparing fragmentation yields per residue between ‘soft’ and ‘harsh’ (see SI Figure 2), the N-terminal region up to residue Glu18 fragments slightly more abundantly in the ‘soft’ condition. Between residue Lys33 and Tyr59, yield is significantly enhanced after initial unfolding of the precursor ion.
Generally, the fragments we observe agree with the higher intensity cleavage sites found in previous UVPD studies obtained at 193 nm . The lesser extent of backbone cleavage we obtain is most likely due to the activation wavelength and therefore both a power and absorption cross-section difference.
Using UVPD post-ion mobility, Warnke et al. demonstrated that 193 nm UVPD is sensitive to underlying structural motifs within [M + 7H]7+ ubiquitin . Our spectra of [M + 6H]6+ contains six out of the eight fragment motifs which were used to assign related conformers (y39, y40, y58, a36, b36 and a37). The two absent ones are higher charged species of two of these fragment ions most likely absent due to the lower charge state of the precursor ion here. Whilst we observe a general increase in the abundance of the mentioned fragmentation motifs when the ATD is shifted towards higher drift times, this does not appear to have been observed for [M + 7H]7+when selectively irradiating portions of the ATD .
To ensure that the increase in fragmentation yield between ‘soft’ and ‘harsh’ conditions was a result of a change in conformation and not a result of increased internal energy from raising the cone voltage, UVPD spectra were obtained for [M + 6H]6+ at each cone voltage between 10 and 120 V in increments of 10 V (SI Figure 3). Changes in fragmentation yield were only observed where changes in cone voltage resulted in a change in arrival time distribution, indicating that the conformational change is the driving factor for fragmentation yield of [M + 6H]6+. In further support of this hypothesis, for [M + 8H]8+ sprayed from 50% methanol, we observed no change in conformation as a function of cone voltage and correspondingly no change in the UVPD fragmentation yield (SI Figure 4). Under the same instrumental conditions, the fragmentation yield does increase as a function of charge state (SI Figure 5) as noted by Brunet et al. for small negatively charged protein ions using photons in the VUV range  and by Cammarata and Brodbelt for positively charged myoglobin using 193 nm .
Since we observe an increase in fragmentation yield as a function of unfolding, the question arises whether the fragments seen under ‘soft’ conditions come from the initial compact conformational families or the more extended conformational family already present in lower abundance. Comparison of the arrival time distribution of the precursor ion from laser off to laser on (SI Figure 6) shows a proportionally larger decrease in intensity of the more extended conformational families as opposed to a uniform decrease across all species, indicating that upon exposure to 213 nm UV photons, more extended conformations of [M + 6H]6+ give up fragments more readily. This indicates that the fragments observed at cone voltage 10 V may arise predominantly from direct UVPD from extended structures present under non-activating conditions.
We compared the arrival times of fragment ions which were observed in both ‘soft’ and ‘harsh’ conditions (SI Figure 7). There was no difference in arrival times of the investigated fragment ions between cone 10 and 120 V, implying that fragments from both source conditions do indeed arise from the same extended conformational families. It is also possible that fragment ions rearrange into a stable conformation on a timescale faster than we are able to measure. Further to this, some reports have indicated that lack of a specific UVPD fragment does not necessarily indicate strong non-covalent associations and suggests a combination of direct dissociation coupled with internal conversion (CID like) events [70, 71] however in ECD experiments it is well-established that lack of fragments correlates with strong non-covalent associations [24, 25, 36, 72]. For ubiquitin, as shown above, and as shall be shown for cytochrome c and myoglobin (below), we can interpret the UV-induced fragments in terms of regions of the proteins that are less encumbered by non-covalent associations. This indicates that in our experiment the UVPD fragments are observed when that region of the protein is detached from the native fold.
When sprayed out of 200 mM ammonium acetate, cytochrome c was observed as ions [M + zH]z+ where z = 6, 7 and 8, with [M + 7H]7+ as the dominant species (SI Figure 8). The charge state distribution did not change as a function of cone voltage; however, a slight reduction in the intensity of salt adducts was seen under more activating conditions. [M + 7H]7+was selected for further investigation.
Due to the difference in fragmentation pattern between ‘intermediate’ and ‘harsh’, it is likely that UVPD is probing both the intermediate and extended conformational families simultaneously. A comparison of the ATDs at each CV setting between laser off and on (SI Figure 10) shows no change in ATD at ‘soft’ and ‘intermediate’ conditions, indicating that all conformational families are probed equally. However, at CV = 120 V, the extended conformational family shows a disproportionally large decrease in intensity compared to the remainder of the more compact family. Fragment ion ATDs, as in the case of ubiquitin, did not change between conditions (SI Figure 11), indicating that fragments either refold to the same conformation during trapping or were produced from the same conformational family.
Myoglobin sprayed from 200 mM ammonium acetate was observed in its holo form with the heme group still attached as ions [M + zH]z+ where z = 7, 8 and 9, with [M + 9H]9+as the dominant species (SI Figure 12), which was selected for further analysis.
No difference was seen when comparing the precursor ATD before and after laser irradiation (SI Figure 14); hence, it is highly likely the absence of UVPD fragments under ‘soft’ conditions is a result of the compact conformation being impervious to dissociation due to the high degree of folding and strong stabilising non-covalent interactions. Out of the 21 cleavage sites, 7 are C-terminal to a lysine or glutamic acid residue. The most intense fragments were obtained on the N-terminal side of proline residues, namely a36, b36 and y54. In contrast to cytochrome c, residues in proximity to the heme group were cleaved (see Fig. 4c); however, some heme retention was observed on y66 highlighting the ability of UVPD at 213 nm to yield fragments which still possess non-covalent interactions, and the strength of the interaction between the cofactor and the protein.
The unfolding sequence of myoglobin as established by other techniques shows the F-helix to unfold first, followed by C-, D-, E- and B-helix, then lastly A, G and H [59, 77, 78, 79, 80, 81]. Consistent with the findings by Cammarata and Brodbelt for 193 nm UVPD of [M + 9H]9+ myoglobin , the G and H-helix which have been shown to form the core of the protein as the most stable structures exhibit the least amount of fragmentation with only two cleavage sites in total. In contrast, however, the F-helix only fragmented at a single residue and therefore appeared stable in our experiment. Most fragmentation occurred towards the C-terminus of the A-helix, C-terminus of the B-helix, the CD loop and the D-helix.
Reducing the charge state of myoglobin from 9+ to 8+ eliminated fragmentation in the GH-loop, H helix and D helix in our experiment (SI Figures 15 and 16) as well as reducing overall fragmentation yield as found with ubiquitin. The cleavage sites with the most abundant fragmentation remain similar between [M + 8H]8+ and [M + 9H]9+ (SI Figure 17).
DC is grateful to the Universita’ degli Studi di Roma La Sapienza for funding his stay at the University of Manchester.
This work was supported by BBSRC grants BB/L002655/1, BB/L016486/1 and BB/M011208/1 which is a studentship award to AT and RB with additional financial support from Waters Corp.
- 2.Chapman, H.N., Fromme, P., Barty, A., White, T.A., Kirian, R.A., Aquila, A., Hunter, M.S., Schulz, J., Deponte, D.P., Weierstall, U., Doak, R.B., Maia, F.R.N.C., Martin, A.V., Schlichting, I., Lomb, L., Coppola, N., Shoeman, R.L., Epp, S.W., Hartmann, R., Rolles, D., Rudenko, A., Foucar, L., Kimmel, N., Weidenspointner, G., Holl, P., Liang, M., Barthelmess, M., Caleman, C., Boutet, S., Bogan, M.J., Krzywinski, J., Bostedt, C., Bajt, S., Gumprecht, L., Rudek, B., Erk, B., Schmidt, C., Hömke, A., Reich, C., Pietschner, D., Ströder, L., Hauser, G., Gorke, H., Ullrich, J., Herrmann, S., Schaller, G., Schopper, F., Soltau, H., Kühnel, K.U., Messerschmidt, M., Bozek, J.D., Hau-Riege, S.P., Frank, M., Hampton, C.Y., Sierra, R.G., Starodub, D., Williams, G.J., Hajdu, J., Timneanu, N., Seibert, M.M., Andreasson, J., Rocker, A., Jönsson, O., Svenda, M., Stern, S., Nass, K., Andritschke, R., Schröter, C.D., Krasniqi, F., Bott, M., Schmidt, K.E., Wang, X., Grotjohann, I., Holton, J.M., Barends, T.R.M., Neutze, R., Marchesini, S., Fromme, R., Schorb, S., Rupp, D., Adolph, M., Gorkhover, T., Andersson, I., Hirsemann, H., Potdevin, G., Graafsma, H., Nilsson, B., Spence, J.C.H.: Femtosecond X-ray protein nanocrystallography. Nature. 470, 73–78 (2011)CrossRefGoogle Scholar
- 3.Liang, B., Tamm, L.K.: NMR as a tool to investigate the structure, dynamics and function of membrane proteins. Nat. Struct. Mol. Biol. 23, 568–474 (2016)Google Scholar
- 8.König, I., Zarrine-Afsar, A., Aznauryan, M., Soranno, A., Wunderlich, B., Dingfelder, F., Stüber, J.C., Plückthun, A., Nettels, D., Schuler, B.: Single-molecule spectroscopy of protein conformational dynamics in live eukaryotic cells. Nat. Methods. 12, 773–779 (2015)Google Scholar
- 10.Stuchfield, D., Barran, P.: Unique insights to intrinsically disordered proteins provided by ion mobility mass spectrometry. Curr. Opin. Chem. Biol. 42, 177–185 (2018)Google Scholar
- 23.Polfer, N.C., Haselmann, K.F., Langridge-Smith, P.R.R., Barran, P.E.: Structural investigation of naturally occurring peptides by electron capture dissociation and AMBER force field modelling. Mol. Phys. 103, 11–12, 1481–1489 (2007)Google Scholar
- 46.Fort, K.L., Dyachenko, A., Potel, C.M., Corradini, E., Marino, F., Barendregt, A., Makarov, A.A., Scheltema, R.A., Heck, A.J.R.: Implementation of ultraviolet photodissociation on a benchtop Q exactive mass spectrometer and its application to phosphoproteomics. Anal. Chem. 88, 2303–2310 (2016)CrossRefGoogle Scholar
- 64.Migas, L.G., France, A.P., Bellina, B., Barran, P.E.: ORIGAMI: a software suite for activated ion mobility mass spectrometry (aIM-MS) applied to multimeric protein assemblies. Int. J. Mass Spectrom. 427, 20–28 (2018)Google Scholar
- 80.Lin, X., Zhao, W., Wang, X.: Characterization of conformational changes and noncovalent complexes of myoglobin by electrospray ionization mass spectrometry, circular dichroism and fluorescence spectroscopy. J. Mass Spectrom. 45, 618–626 (2010)Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.