Analysis of SUMO-Isopeptides with Typical Full-Length Tryptic Iso-chains
Using trypsin for proteolytic enzyme digestion of a protein that has been SUMO(1)ylated, to generate a SUMO(1)-isopeptide results in typical C-terminal cleavage of the lysine residue at position 78, resulting in the generation of a full- length tryptic iso-chain with a length of 19 amino acids. Typical tryptic cleavage C-terminal to the arginine at position 61 of the SUMO(2) iso-chain and position 60 of the SUMO(3), resulting in the generation of SUMO(2)/(3)-isopeptides containing a full-length tryptic iso-chain of 32 amino acids. The iso-chains generated on SUMO-isopeptides from typical tryptic cleavage are much longer than the iso-chains generated in Ub-isopeptides (discussed later). The SUMO(1)- isopeptides and SUMO-(2/3)-isopeptides derived from tryptic digestion contain full-length iso-chains consisting of 19 (ELGMEEEDVIEVYQEQTGG) and 32 (FDGQPINETDTPAQLEMEDEDTIDVFQQQTGG) amino acids, respectively. These large iso-chains are highly charged, therefore rendering the SUMO- isopeptide a highly charged species for electrospray-based MS analysis. The MS-based approaches to analyze these highly charged SUMO-isopeptides with these full-length tryptic iso-chains provide strategies that have been developed and applied to improve the analytical performance of these analytes in targeted approaches.
Hybrid linear ion-trap Fourier transform-based mass spectrometry with low-energy ion-trap CID has been coupled to liquid chromatography via nano-electrospray sources to analyze highly charged SUMO-isopeptides containing a full-length tryptic iso-chain in a targeted approach involving two parts . The first part of this work flow utilized the high-resolution mass spectrometer and high mass accuracy measurement capabilities of Orbitrap and FT-ICR mass spectrometers to identify their naturally higher precursor ion charge states of a set of full-length SUMO(1)-isopeptides and SUMO(2)- isopeptides generated from independent post-tryptic and Lys-C digestion of polymeric-SUMO protein chains produced in in vitro SUMOylation assays. This set of SUMO-isopeptides was then analyzed by low-energy CID in the ion trap and the m/z of the predominantly multiply charged product ions were measured in only the ion trap when the LTQ-FT-ICR was used or in both the ion trap and the Orbitrap when using the LTQ Orbitrap. Manual analysis of the highly complex CID MS/MS spectra was difficult due to the domination of multiply charged b'- and/or y'-type product ions characteristically generated under low-energy CID conditions from the SUMO-isopeptides with highly charged full-length tryptic SUMO iso-chain. (Note: b and y product ions from the iso-chain and backbone of the SUMO-isopeptides are distinguished as follows: b'-/y'-type and c'- and z'-type product ions refer to ions from the isochain of SUMO- and Ub-isopeptides and b-/y-type and c- and z-type product ions refer to ions from the backbone of SUMO- and Ub-isopeptides). Consequently, there were a limited number of product ions identified from the peptide backbone.
In order to analyze the resulting CID MS/MS spectra of these highly charged SUMO-isopeptides, the authors used a specialist software, which transformed them into a “virtual” SUMO-isopeptide, and subsequent in silico fragmentation enabled the product ions of the iso-chain to be calculated and identified on the CID MS/MS spectra. The second part of the approach involved using the analytical information such as the retention times, their naturally higher precursor ion charge states, and their fragmentation behavior to successfully target and identify the same set of SUMO-isopeptides in a more complex sample from cultured mammalian cells. The specialist software and other available software [34, 35] used to assist with the interpretation of the complex SUMO-isopeptide CID MS/MS spectra suffer from:
- 1) An input of poor spectral quality populated by dominant multiply charged product ions from the SUMO-isopeptide's highly charged iso-chain, further suffering from limited sequence coverage of product ions generated from the isopeptide backbone, thereby limiting the backbone's comprehensive structural elucidation and identification
- 2) Limited consideration for the impact of multiple variable modifications that may occur along the iso-chain, such as multiple events of deamidation from the presence of two to five glutamine residues.
Another targeted approach used to analyze full-length tryptic SUMO-isopeptides utilized the capability of specialized Fourier transform ion cyclotron resonance (FT-ICR)-based mass spectrometry to provide high-resolution mass and high mass accuracy measurements. In addition, FT-ICR is capable of utilizing a range of specialized low-energy-based fragmentation techniques: electron capture dissociation (ECD), associated activated-ion electron capture dissociation (AI-ECD), and infrared multiphoton dissociation (IRMPD) in MS/ MS mode. ESI-FT-ICR with ECD and IRMPD has been used to directly analyze a tryptic digest containing SUMO(1)-isopeptides bearing full-length tryptic SUMO-1 iso-chains, SUMO(1)ylated RanGaplprotein fragment, and SUMO(1) ylated RanB2 protein fragments . The IRMPD MS/MS spectrum generated from the analysis of the SUMO(1)-isopeptide derived from the SUMO(1)ylated RanGapl generated a series of only singly charged b/b'- and y/y'-type product ions (b'/y' iso-chain ions labeled with b/y (S) in Figure 6.1a) from the backbone and iso-chain of the isopeptide (Figure 6.1a). The generation of only singly charged products greatly reduces the complexity of the IRMPD MS/MS spectrum. This is in contrast to the predominance of multiply charged product ions observed in CID MS/MS spectra, which are typically generated from the isochain of SUMO(1)/(2)-isopeptides. The shift from the generation of multiply charged to only singly charged product ions is indicative of IRMPD MS/MS spectra . Although IRMPD fragmentation of the precursor ion is less efficient than CID fragmentation, the reduction in complexity observed in the IRMPD MS/MS spectrum of the SUMO(1)-isopeptide is advantageous and enables improved structural elucidation of both the backbone and the isochain. The additional series of predominantly b'- and y'-type-related neutral loss ions generated from the neutral loss of H2O from SUMO(1) iso-chain of the isopeptide can be attributed to secondary dissociation events of either initial b'- and y'-type product ions or the dehydrated precursor ion, typically indicative of the IRMPD fragmentation technique and resulting IRMPD MS/ MS spectra of peptides . The generation of these product ions arising from
Figure 6.1 (a) IRMPD MS/MS spectrum of a SUMO(1)-isopeptide tryptically derived from a SUMO(1)ylated RanGAP1418-587 protein fragment. (b) An ECD MS/MS of a SUMO(1)- isopeptide tryptically derived from a SUMO(1)ylated RanGAP1418-587 protein fragment.
(S) refers to product ions generated from the SUMO iso-chain. v=harmonic and (A) = artifact not removed in quadrupole/SWIFT isolation. Source: Cooper 2005, . Reproduced with permission from American Chemical Society.
the neutral loss could also be favored due to the presence of multiple glutamic acids within the SUMO(1) iso-chain. This could account for their underfragmentation at the N-terminal position of these types of related y'-ions and the C-terminal position of b'-ions generated from the SUMO(1) iso-chain of the isopeptide, which is a fragmentation behavior observed under low-energy CID conditions of peptides . The complementary ECD MS/MS spectrum of the same SUMO(1)-isopeptide generated a series of both singly charged c- and z-type product ions from the backbone of the isopeptide (Figure 6.1).
The characteristic lower fragmentation efficiency of ECD (also observed with its analogous fragmentation technique, ETD) compromises the abundance of product ions; however, the characteristic sequence-independent fragmentation pattern that occurs under ECD fragmentation of peptides can reduce the complexity of the spectra, enabling the c- and z-type product ions to be identified on the ECD MS/MS spectrum of the SUMO(1)-isopeptide backbone. By contrast, only the generation of 1 c'-type product ion (note: c'-type product ion from iso-chain labeled as c (S) in Figure 6.2b) from the iso-chain of the isopeptide was observed. It was suggested that this was most likely due to the predominantly acidic physicochemical nature of the iso-chain and the lack of a basic residue to enable efficient ECD fragmentation to occur. Limited fragmentation under AI-ECD conditions also occurred along the iso-chain of the SUMO(1)-isopeptide generated from a SUMO(1)ylated RanB2 protein fragment. The limited fragmentation of the SUMO(1) iso-chain under ECD conditions could indicate that the physicochemical nature of the full-length SUMO(1) iso-chains may prevent SUMO(1)-isopeptides from being amenable
Figure 6.2 An example of a collision-cell CID MS/MS spectrum of a SUMO(2)-isopeptide atypically tryptically derived from a SUMO(2)ylated SUMO(2) protein. Source: Chicooree, 2013 . Reproduced with permission of Wiley.
to full sequence structural elucidation by ECD. However, the resulting few isochain ions that are generated from limited ECD fragmentation of the iso-chain could be used as iso-chain diagnostic ions. This provides scope to observe isochain-specific diagnostic fragmentation patterns and subsequently characterize product ions, provided the MS/MS spectra are reproducible. Although these specific fragmentation techniques and the associated MS instrumentation are specialized, the limited complexity and additional improvement in production coverage of the SUMO-isopeptide backbone that is observed in the resulting MS/MS spectra using these ECD and IRMPD to complement each other is a contrast to the high complexity of those observed from CID collision-cell and trap-type analysis of SUMO(1/2)-isopeptide also bearing full-length tryptic isochains . The reduced complexity of the product-ion spectra benefits de novo sequencing and enables easier interpretation, structural elucidation, and assignment of the product ions by comparison with in silico digestion of the SUMO(1)-isopeptides without the need to employ specialist software.
Another approach has been developed using a combination of QTOF-based nanoLC-nESI-MS/MS and MALDI-TOF/TOF MS/MS analysis to analyze SUMO-isopeptides bearing full-length tryptic iso-chains. nanoLC-nESI-MS/ MS was used to utilize its high resolution and mass accuracy in order to screen for potential SUMO-isopeptide ions based on the presence of quadruply charged monoisotopic signal clusters from a tryptic digest sample generated from SUMOylated proteins produced in vitro . These proteins included a Ubc 9 protein that had been SUMOylated by a SUMO(2) protein and human centromere protein C10 and C28 fragments that had been SUMOylated by SUMO(1) and SUMO(2) proteins. The potential SUMO-isopeptide ions were targeted for analysis using MALDI-TOF/TOF MS/MS in order to utilize its ability to induce characteristic fragmentation patterns relating to preferential C-terminal cleavage of aspartic acid residues and generate singly charged product ions from the singly charged SUMO-isopeptide ions. Full-length tryptic SUMO(2) (and SUMO(3)) iso-chains of the singly charged tryptic SUMO(2)- isopeptide ion contain multiple aspartic acid residues and full-length tryptic SUMO(1) iso-chains of the singly charged tryptic SUMO(1)-isopeptides contain a single aspartic acid residue. These aspartic acid residues can often be exploited under MALDI-TOF/TOF high-energy CID by facilitating the fragmentation of a characteristic series of highly abundant b'- and/or y'-type product ions from the iso-chain due to their preferential C-terminal fragmentation.
This type of fragmentation is described as the “aspartic acid effect” and occurs via a charge-remote fragmentation pathway. An in-depth explanation on this fragmentation mechanism and its origins from high-energy CID experiments can be found in Paizs and Suhai (2005) . Although the occurrence of this type of preferential fragmentation and subsequent release of abundant ions prevents complete sequence coverage of each amino acid within the iso-chain or backbone, the aspartic acid residues are distributed along the full-length of the SUMO-(2/3) iso-chain to the extent that their b'- and/or y'-type iso-chain-specific product ions are representative of the large portions of the iso-chain to a good degree whereby structural elucidation can be facilitated. In addition to the generation of these iso-chain-specific product ions, the SUMO(2)-isopeptides analyzed by Chung et al. produced high-energy CID MS/ MS spectra containing both b- and y-type product ions from the isopeptide backbone. Although the SUMO(1) iso-chain contains one aspartic acid residue, the advantageous nature of TOF/TOF high-energy CID MS/MS spectra being highly reproducible  resulted in additional y-type product ions being generated from both the backbone and iso-chain of the SUMO(1)-isopeptides. The generation of characteristic product ions from the iso-chain of all the isopeptides analyzed enabled structural elucidation of the isopeptides from in vitro samples without the need for specialized bioinformatic algorithms. Although the authors indicated that comprehensive bioinformatic algorithms could be developed for global analysis based on the characteristic fragmentation patterns and additional observations of SUMO(2)/(3) iso-chains, they have presented in their targeted approach. The generation of only singly charged product ions under high-energy CID conditions used in MALDI-TOF/TOF is also an advantage in that clean SUMO-isopeptide tandem MS/MS spectra are observed with less complexity in comparison to the SUMO-isopeptide MS/MS spectra produced using low-energy CID, which contain predominantly multiply charged product ions generated from neutral losses and internal ions. However, there is an initial reliance on two instrument platforms to apply this targeted approach.
In summary, the targeted MS approaches discussed have provided strategies to overcome the inherently challenging physicochemical nature of the SUMO iso-chains and improve the overall analytical performance of these challenging analytes. However, all these approaches are reliant on multistage workflows, specialist software programs, specialized fragmentation techniques, and multiinstrument platforms.