Tandem mass spectrometry fragmentation patterns of sulfo-SDA cross-linked peptides

Abstract

Cross-linking mass spectrometry is rapidly becoming a choice method for determining a protein's higher-order structure as well as capturing inter-protein interactions. In particular, diazirene-based photo-activatable cross-linkers, such as sulfo-SDA have been shown to be effective at generating high-density cross-linking data. Previously, we have shown that this method may be used to study binding orientation between two non-covalently linked complexes; however, several unexpected ions were noted in the MS2 spectra. In this study, the tandem mass spectrometry fragmentation patterns of sulfo-SDA-initiated cross-linked peptides under higher-energy collision induced (HCD), collision induced (CID) and electron transfer with supplementary HCD (EThcD) dissociations are discussed. The analysis revealed substantial insights into localising cross-linking sites, which is essential for accurate determination of protein higher-order structural characteristics.

Introduction

Cross-linking mass spectrometry (XL-MS) is an emerging technique used in capturing both non-covalent intermolecular interactions and protein assembly in native environments. This versatility for higher-order structural analysis has ensured XL-MS has become an important part of the structural biology toolkit over recent years.^1,2

Heterobifunctional cross-linkers allow for sequential reactions with two functional groups. In these methods, covalent bonds are formed between two amino acids within a given proximity. Proteins are subsequently proteolytically digested and peptides from bonded residues can be sequenced by liquid chromatography–tandem mass spectrometry (LC-MS/MS). ³

In particular, diazarene-based photo-crosslinkers have recently been developed. Sulfosuccinimidyl 4,4′-azipentanoate (sulfo-SDA) is a popular example and has been shown to be effective at generating high-density cross-linking data. ⁴ Whilst the use of this reagent is well reported in the literature, individual MS/MS data are not always discussed in detail and often a significant portion of fragment ions remain unassigned.^4,5

Figure 1 shows the mechanism by which sulfo-SDA cross-links are initiated. In the first stage of cross-linking, the amino group reacts with a lysine side chain or protein N-terminus (and to a lesser extent, serine, threonine or tyrosine side chains). Upon exposure to UV light at 365 nm, the diazirene moiety forms a highly reactive carbene intermediate that has the capacity to react with either the side chain or backbone carbon of any amino acid.

Figure 1.

Mechanism of formation of sulfo-SDA cross-link formation. Red dashed line shows additional fragmentation observed during higher-energy collisional dissociation (HCD), collision induced (CID) and electron transfer with supplementary HCD dissociation (EThcD). Green dashed line shows additional fragmentation observed during EhcTD.

In our previous study, we showed this cross-linker was effective at determining the binding orientation of different T-cell receptors when binding to peptide human leukocyte antigens. ⁶ However, the mass spectrometry data provided several interesting insights into the behaviour of sulfo-SDA-initiated cross-links under higher-energy collisional dissociation (HCD). Many of the most abundant peaks in the MS2 spectra would have remained unassigned if the cross link was fully maintained under MS/MS fragmentation. It was of particular interest that we observed an additional fragmentation under HCD between the two atoms labelled carbon X and carbon Y (see Figure 1). This was unexpected as HCD typically exclusively cleaves C-N bonds.

In this letter, we thoroughly discuss the fragmentation of this cross-linker under three different MS2 techniques: HCD, collision induced (CID) as well as electron transfer with supplementary HCD (EThcD). This will enhance the mass spectrometry community's ability to accurately localise cross-links to specific residues. Full determination of cross-linking sites will enable a more precise understanding of binding interfaces.

Methods

Two hundered micrograms human serum albumin (HSA) was crosslinked with sulfo-SDA, using a protein-to-crosslinker molar ratio of 1:200. Samples were photoactivated using ultraviolet light irradiation at 365 nm for 50 min at 200,000 μJ/cm2 using a UVP CL-1000 UV crosslinker (Akribis, Cheshire, UK).

Next, samples were denatured with guanidine hydrochloride, reduced with dithiothreitol, alkylated with iodoacetamide and tryptically digested. Peptides were separated using online RP-LC and directly eluted via nESI onto an Orbitrap Fusion (Thermo Fisher Scientific) and analysed via MS/MS using HCD, CID or EThcD. Cross-linked peptides were identified using XiSearch. ⁷ The raw data for positive identifications was manually interrogated.

Results and discussion

A model protein (HSA) was cross-linked using sulfo-SDA. Subsequently, HSA was tryptically digested, separated over an LC gradient and analysed via MS/MS using HCD, CID and EThcD.

Figure 2 shows the HCD mass spectra of a representative pair of cross-linked peptides that were identified in each of the three datasets. A full list of identified ions is shown in Supplementary Table 1. The cross link was shown to be between peptides TCVADESAENCDK and FKDLGEENFK. For the remainder of the letter, these shall be referred to as Peptide 1 and Peptide 2 for simplicity.

Figure 2.

Higher-energy collisional dissociation (HCD) tandem mass spectrometry (MS/MS) for sulfo-SDA cross-linked peptides 1 and 2. b and y ions are shown in red or blue dependent on their corresponding peptide. Ions resulting from the additional cleavage adjacent to the SDA cross-linker are shown in dark red or dark blue. Some less abundant ions are not labelled in the spectra. A full list of assignments is included in the relevant supplemental tables.

Whilst working with the assumption that the sulfo-SDA cross link should be fully maintained under HCD, b and y ions were assigned for both peptides using XiSearch and verified. Additionally, unassigned fragment ions in the MS2 spectra missed by XiSearch were manually interrogated.

Comprehensive sequence coverage for Peptide 1 (without peptide 2 linked) was noted across the experimental b and y ions. This is of interest, as if the cross link was fully maintained during HCD, not all these fragmentation assignments should be possible.

Next, we searched for fragment ions on Peptide 1 which contained the additional mass of the cross-linker and Peptide 2. These ions should theoretically exist on fragments including and following the amino acid where the cross-linker has linked. Ions y₈-y₁₁ plus the mass of sulfo-SDA and Peptide 2, in the 2+ charge state were noted, suggesting the cross-linking site is E6 in Peptide 1.

The presence of y₉-y₁₁ without the partner peptide would be impossible for this assignment if the cross-linker was fully maintained under HCD, as well as b₆ and b₈-b₁₂.

B and y ions were similarly interrogated for Peptide 2. A prominent y ion series was observed, with y₁-y₄ and y₆-y₈ reported. Y₉ (without the partner peptide) at m/z ∼1079.54 was not present, though this is expected as sulfo-SDA will typically bind to lysine residues in at least one of the cross-linked peptides. This is further supported by the missed tryptic cleavage site, which is generally, though not always, observed on a binding residue. Interestingly, there was not a fragment corresponding to the mass of y₉ plus the mass of the cross-linker and Peptide 1, in any charge state, which would be expected if the cross link was maintained under HCD.

However, an ion was observed at m/z 1161.5797. This was hypothesised to correspond to y₉ plus the mass of the cross-linker only. This is supported by the comprehensive b ion series where each ion of b₂-b₉ was noted with the additional mass of the cross-linker. Moreover, an ion corresponding to the mass of Peptide 1 as well as an ion corresponding to Peptide 2 plus the mass of sulfo-SDA are noted in the MS/MS spectrum.

Similarly to Peptide 1, it would be impossible to localise the cross-linking site in Peptide 2 without first identifying these additional fragmentation patterns. Whilst K2 is the most likely cross-linking site, it is also possible in theory for K10 to be incorrectly assigned as the binding site. This would severely alter the conclusions of any data analysis investigating binding interfaces.

Whilst some cross-linkers are designed to be MS cleavable, with the aim of generating diagnostic fragment ions, ⁸ sulfo-SDA has not been manufactured with this in mind. Indeed, sulfo-SDA has been reported to lack MS-cleavable functional groups, which prevents confident identification of certain cross-linked peptides. ⁹ Whilst HCD is typically the choice method for the identification of cross-links, CID and EThcD fragmentation methods may also be used.

Figure 3 shows the CID spectra for the same cross-linked peptide (see also Supplementary Table 2). Here, the base peak in the mass spectrum corresponds to [Pep2]²⁺ with the additional mass of the cross linker. The second most abundant peak is [Pep1]⁺. These data are similar to the HCD data in that a major fragmentation event during CID is between Carbon X and Carbon Y noted in Figure 1.

Figure 3.

Collision-induced (CID) tandem mass spectrometry (MS/MS) for sulfo-SDA cross-linked peptides 1 and 2. b and y ions are shown in red or blue dependent on their corresponding peptide. Ions resulting from the additional cleavage adjacent to the SDA cross-linker are shown in dark red or dark blue. Some less abundant ions are not labelled in the spectra. A full list of assignments is included in the relevant supplemental tables.

This mechanism is not consistent with the mobile proton model which describes peptide fragmentation under both CID and HCD. ¹⁰ The model describes the energy imparted in the collision results in proton transfer from basic amino acid side chains or the N-terminus to amide nitrogen atoms resulting in a C-N breakage, which are typically the weakest bonds. It is therefore unusual that a cleavage of the C-C bond is noted here.

The same peptide was similarly selected for fragmentation by EThcD (see Figure 4 and Supplementary Table 3 for full ion assignments). In contrast to HCD and CID, the cross-linker is largely maintained during this method of fragmentation. The exceptions to this are ions which correspond to the masses of Peptide 1, Peptide 2 as well as Peptide 2 plus the additional mass of the cross-linker. This suggests that in addition to the C-C cleavage previously described, there is a C-N cleavage that also occurs during this fragmentation (see Figure 1). These ions are low in relative abundance, suggesting this is not a major fragmentation pathway.

Figure 4.

Electron transfer with supplementary HCD dissociation (EThcD) tandem mass spectrometry (MS/MS) for sulfo-SDA cross-linked peptides 1 and 2. b and y ions are shown in red or blue dependent on their corresponding peptide. Ions resulting from the additional cleavage adjacent to the SDA cross-linker are shown in dark red or dark blue. Some less abundant ions are not labelled in the spectra. A full list of assignments is included in the relevant supplemental tables.

Conclusion

In summary, our data shows the susceptibility of the cross-linker to be cleaved under HCD and CID techniques. Indeed, if one was to assume the cross-linker was fully maintained under these conditions, as expected, it would be impossible to accurately determine cross-linking sites. The data presented here further clarifies the different fragmentation pathways of sulfo-SDA peptides, ultimately leading to better fragment assignments in a technique where identifying binding sites to single residue specificity is critical.