Characterisation of the interaction between arginine methyltransferase

Subscriber access provided by The University of British Columbia Library

Article

Characterisation of the interaction between arginine methyltransferase Hmt1 and its substrate Npl3: use of multiple crosslinkers, mass spectrometry approaches and software platforms Daniela-Lee Smith, Michael Goetze, Gene Hart-Smith, Marc R. Wilkins, and Tara Bartolec Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b01525 • Publication Date (Web): 13 Jul 2018 Downloaded from http://pubs.acs.org on July 16, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Characterisation of the interaction between arginine methyltransferase Hmt1 and its substrate Npl3: use of multiple crosslinkers, mass spectrometry approaches and software platforms Daniela-Lee Smith1, Michael Götze2, Tara K. Bartolec1, Gene Hart-Smith1, Marc R. Wilkins1 1

Systems Biology Initiative, School of Biotechnology and Biomolecular Sciences,

UNSW Sydney, NSW 2052, Australia 2

Institute of Biochemistry, Martin Luther University Halle-Wittenberg,

Kurt-Mothes-Str. 3, D-06120 Halle (Saale), Germany

Corresponding Author: Prof. Marc Wilkins E: [email protected] T: +61-2-9385-3633

ACS Paragon Plus Environment

1

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

Abstract This study investigated the enzyme-substrate interaction between the Saccharomyces cerevisiae arginine methyltransferase Hmt1p and nucleolar protein Npl3p, using chemical crosslinking-mass spectrometry (XL-MS). We show that XL-MS can capture transient inter-protein interactions that occur during the process of methylation, involving a disordered region in Npl3p with tandem SRGG repeats, and confirm that Hmt1p and Npl3p exist as homo-multimers. Additionally, the study investigated the interdependencies between variables of an XL-MS experiment, that lead to the identification of identical or different crosslinked peptides. We report that there are substantial benefits, in terms of biologically relevant crosslinks identified, that result from the use of two mass-spectrometry cleavable crosslinkers disuccinimidyl sulfoxide (DSSO) and disuccinimidyl dibutyric urea (DSBU), two fragmentation approaches CID+ETD or SteppedHCD, and the two programs MeroX or XlinkX. We also show that there are specific combinations of XL-MS methods that are more successful than others for the two proteins investigated here; these are explored in detail in the text. Data are available via ProteomeXchange with identifier PXD008348.


2

Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Interactions between proteins are essential for almost all cellular functions. The study of protein-protein interactions (PPIs) can thus provide detailed insights into cellular organization and regulation. On a large scale, studies of PPIs can also generate protein interaction networks and systems-level views of the cell 1. To date, approaches involving two-hybrid techniques or affinity purification-mass spectrometry (AP-MS) have been widely used2-3, yet they do not readily provide any direct insights into where and how proteins actually interact. An alternative method to study PPIs is crosslinking mass spectrometry 4. Crosslinking mass spectrometry (XL-MS) combines chemical crosslinkers, protease digestion and tandem mass spectrometry, whereby fragmentation of the crosslinked peptides is used to discover intra- and inter-protein crosslinks5.

Early XL-MS studies used non-MS cleavable crosslinkers such as BS3, BS2 G and DTTSP 6-10. The masses of crosslinked peptides could be measured but it was difficult to identify the peptides themselves. This was due, in part, to the large search space that exists when searching for two crosslinked peptides (the “N-square” problem11) and low confidence identification that can arise from co-fragmentation of crosslinked peptides. Nevertheless, non-MS cleavable crosslinkers have been successfully used for the study of many large complexes (e.g. the mediator/Pol II preinitiation complex) and for the proteome-scale study of Chaetomium thermophilum12,13.

New MS-cleavable linkers, such as DSSO5, DSBU14 and Protein Interaction Reporter (PIR)15, have helped address many of the prior limitations of XL-MS. They allow predictable generation of high-intensity reporter ions. In conjunction with mass


3


Page 4 of 27

spectrometers such as the Q-Exactive and the Fusion Tribrid, this permits the high confidence identification of crosslinked peptides from simple or more complex samples. New software, including XlinkX 11, 16 and MeroX17, can take advantage of these reporter ions and peptide fragmentation data, circumventing the “N-square” problem11.

Three new-generation XL-MS pipelines, which combine a crosslinker and software tool, are DSBU/MeroX 14, 17, DSSO/XlinkX 5, 11, 16and PIR/ReAct 15, 18. Each pipeline has been applied to different biological systems, both small and large scale, using crosslinkers and fragmentation strategies which maximize the number of identifications. MeroX17, paired with DSBU and native MS, has produced new models of the intrinsically disordered protein p5319-21 and insights into the Ca2+ dependent binding structure of GCAP-217 . XlinkX11, 16, which was originally developed for use of DSSO with CID+ETD-based MS2 analysis, has recently improved its capacity to identify crosslinks by use of MS3 data16. By analysing complex samples with a hybrid MS2-MS3 CID+ETD approach, Heck and coworkers reported 1,158 and 3,301 crosslinks from E. coli and HeLa lysates at 1% FDR respectively.

Here we examined three variables of an XL-MS experiment - crosslinker, instrument and fragmentation, and program. We sought to understand how each variable contributed to differences in identifications and how combinations of variables could maximise biological insight. Specifically, we evaluated the XlinkX 2.0 and MeroX platforms, both which can be performed on standard Q-Exactive and Fusion Lumos Tribrid benchtop instruments, using commercially available DSSO or DSBU


4

Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


crosslinkers. The interaction between two transiently interacting Saccharomyces cerevisiae proteins was examined, histone arginine methyltransferase 1 (Hmt1p) and one of its substrates heteroribonucleolar protein Npl3p. This revealed that crosslinker, fragmentation and algorithm all impact on the number and type of crosslinks discovered. We conclude with four recommendations to aid in the choice of XL-MS strategies for similar experiments.

Methods In-vitro crosslinking with DSSO or DSBU - Recombinant Npl3p and Hmt1p were expressed and purified according to 22. Crosslinking reactions were undertaken in accordance with previous studies 11, 20. Specifically, equimolar (~5 µM) amounts of Npl3p and Hmt1p were mixed in 20mM HEPES (pH 8), 150mM NaCl, 0.5mM DTT and 1mM of NHS ester-based crosslinker: disuccinimidyl sulfoxide (DSSO, a gift from Albert Heck) or disuccinimidyl dibutyrin urea (DSBU). After 5 min at 21 °C, AdoMet was added to 50uM and the mixture incubated for 1h at 21 °C. Crosslinked proteins were reduced, alkylated digested with Trypsin (Promega) as per 11 then desalted using C18 Sep-Pak (Waters) cartridges according to 11. Elutions were dried by SpeedVac and resuspended in 0.1% (v/v) formic acid.

LC-MS/MS analysis of peptides – Tryptic peptides were separated on an Ultimate 23

3000 UHPLC nano system (Thermo Fischer Scientific)

with solvent A (0.1% v/v

formic acid in 2% v/v CH3CN) and solvent B 0.1% (v/v) formic acid in 45% (v/v) CH3CN. A 4 min 0.1% trifluoroacetic acid wash was followed by 5-160 min linear gradient from 0% B to 45% B, 160-164 min, linear gradient to 80% B, 164-165 min 80% B, 165-179 min linear gradient to 2% B. Each sample was analysed on a Tribrid


5


Page 6 of 27

Fusion Lumos mass spectrometer and, for the same sample, also on a Q-Exactive Plus. The Tribrid Fusion Lumos was used for MS2 CID+ETD fragmentation with parameters adapted from 16 and the Q-Exactive Plus was used for Stepped HCD analysis. The analyses were undertaken in technical duplicate. Full methods for both instruments are given in Supplementary Information.

Data analysis and software parameters – Raw data was converted to .mgf as detailed in Supplementary Information. Mascot searches for protein methylation on Npl3 were performed according to 24. Crosslinks were searched for with MeroX 1.6.6 and XlinkX 2.0, using the same parameters where possible; details of this are given in the Supplementary Information. Each file was analysed with two crosslinking 11, 17

identification algorithms: Reporter-ion algorithms and Precursor algorithms

.

FDRs for MeroX and XlinkX were set at 1%. For the biological investigation of the Npl3p/Hmt1p interaction, MeroX searching was done using KSTY and the Nterminus as crosslinkable entities and results were combined with those from XlinkX (which only considers K-K crosslinks). For all technical comparisons, a second MeroX search was performed with K and the N-terminus as crosslinkable entities, to allow direct comparison with the results from XlinkX. The only peptides considered in technical comparisons were those that were identified as exact duplicates (same crosslinker, same peptides, same sites of crosslinking).

Results and Discussion One aim of this study was to generate the maximum biological insight into a crosslinked protein sample. To do this, the effects of three variables of an XL-MS experiment, crosslinker, fragmentation strategy and analysis software were studied in


6

Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


context of one another (Figure S-1). and the results of their use in different combinations were compared. S. cerevisiae arginine methyltransferase Hmt1p and its substrate protein Npl3p were chosen for analysis. This enzyme/substrate system is an excellent case study for a number of reasons. It contains a series of stable interactions in that Hmt1p forms higher order multimers 25, and there is evidence Npl3p forms dimers 26-28 This system also contains a transient interaction between ordered and disordered regions, in that Npl3p has 17 arginines within its intrinsically disordered SRGG region that are subject to methylation by Hmt1p 24. Notably, this will be the first XL-MS characterization of a protein methyltransferase enzymatically modifying its substrate. Our experimental design is unique in that it allows the study of stable, transient and intrinsically disordered protein-protein interactions within one XL-MS experiment.

XL-MS captures stable and transient interactions between Hmt1p and Npl3p including those in disordered regions We first confirmed that Hmt1p had methylated, and was thus interacting with, its substrate Npl3p under crosslinking conditions. We found four monomethyl-, three dimethyl- and four methyl/dimethyl-arginines on the SRGG region of Npl3p (see Table S-1), all of which confirmed known Hmt1p methylation sites24. We next collated crosslinking results, found in at least technical duplicate, from all crosslinkers and mass spectrometry techniques. A total of 70 non-redundant crosslinks were identified by MeroX and/or XlinkX (Figure 1A, Table S-2 and S-3). Of these, 8 Hmt1p-Npl3p interprotein crosslinks were identified by MeroX (Figure 1A, Table S-2). XlinkX did not identify any interprotein crosslinks, only intraprotein crosslinks in Hmt1p (Table S-3). This difference is due to MeroX having a capacity


7


Page 8 of 27

to identify crosslinks involving serines, but XlinkX being constrained to the identification of crosslinks involving lysines. Biological replication of the experiment confirmed Hmt1p intraprotein crosslinks, interactions between Hmt1p and multiple sites in the SRGG region of Npl3p, and intraprotein crosslinks in this SRGG region (Figure S-2).

Figure 1: Inter- and intraprotein crosslinks between methyltransferase Htm1p and its substrate protein Npl3, identified by XL-MS. A) All crosslinks identified by MeroX and/or XlinkX at FDR 1% 29. Proteins Hmt1p and Npl3p are represented linearly; blue block represents the intrinsically disordered SRGG region in Npl3 where methylation occurs. Green lines represent interprotein crosslinks between Hmt1p and Npl3p, purple semicircles represent intraprotein crosslinks whereas small red loops represent crosslinks between the same peptide, indicating homomultimerisation. B) Homodimer of Hmt1p (PDB structure 1G6Q) with all intraprotein crosslinks shown. Purple lines indicate a distance measure between the α-carbons (pink spheres) of crosslinked residues, and match the purple lines for Hmt1p in panel A. Distances between α-carbons are shown in Ångstroms, on one monomer (grey) of Hmt1p. The other monomer of the dimer is shown in brown.


8

Page 9 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Of the 8 Hmt1p-Npl3p crosslinks, we found 6 interprotein links between Hmt1p and the SRGG region of Npl3p, where arginine methylation occurs24, 30. The most prominent hotspot of Hmt1p, which bound to the SRGG region, involved K246. Other interprotein interaction sites on Hmt1p involved amino acids towards its Cterminus (K322 and K344) and in the SAM-binding site (T110). The position of these crosslinks, and their importance for Hmt1p-Npl3p interactions, corroborates findings from the literature. The deletion of the 21 C-terminal amino acids of Hmt1p (residues 327 to 348) completely abolishes the dimerization of Hmt1p and its capacity to bind to Npl3p 25. This deletion includes D333, which in the hexameric form of Hmt1p likely forms a salt bridge with K246, in the tip of a loop in a neighboring molecule of Hmt1p. Mutations of amino acids in the SAM binding pocket of Htm1p and its active site (e.g. G68R) are also known to be detrimental for the binding to Npl3p 25, 31. It is notable that our experiment is the first direct observation of Npl3p contacting its methyltransferase Hmt1p and captures a transient interaction during the methylation process. It shows that Hmt1p can bind Npl3p in many places, which is consistent with the methylation found in 17 SRGG motifs in an extended disordered region 24, 30. Similar types of interactions will likely occur between Htm1p and its other SRGGcontaining substrates, such as Gar1p and Nop1p. Whilst outside the scope of this study, iterative mutagenesis and crosslinking experiments (as done by Liu and colleagues 32) could further our understanding of Hmt1p interactions and activity.

In addition to interprotein crosslinks, our XL-MS analysis of Hmt1p and Npl3p revealed a large number of intraprotein crosslinks and evidence of homo-multimers. MeroX identified 58 intraprotein crosslinks (Figure 1A, Table S-2). Of these, three were found to occur between exactly the same residues of Hmt1, and two were


9


Page 10 of 27

between exactly the same residues of Npl3, providing evidence that Hmt1p and Npl3p each form homodimers. XlinkX identified 21 intraprotein crosslinks involving Hmt1. Two of these were evidence of homodimer formation (Table S-3) and confirmed the same intraprotein links found by MeroX, being K246-K246 and K340K340. To assess the consistency of Hmt1p intraprotein crosslinks with structure, relevant links were mapped to PDB entry 1G6Q using Pymol and C-alpha to C-alpha distances were measured. All were within the 25 Angstroms distance constraints of the linkers 5, 14 when mapped onto the Hmt1p dimer (Figure 1B) or homo-hexamer (Figure S-3). The only exception to this was the K340-K340 link which measured a distance of 40.3 Angstroms; this is likely due to flexibility in the homohexamer. For Npl3p, its intraprotein crosslinks and multimerisation status could not be evaluated in a similar manner as there is no known 3-D structure. However affinity purification mass spectrometry 26, two hybrid 22, 28 and mutation analysis 27 has shown Npl3p to homo-dimerise and that its SRGG region is essential for self-association. MeroX crosslinks S349-S349 and S343-S343 are consistent with this and it is likely that other Npl3p intraprotein crosslinks in the SRGG region arose from self-association. Expectedly, we did not find any inter- or intraprotein crosslinks in the N-terminal third of Npl3p. This reflects the complete lack of lysines in this region (amino acids 1 to 150).

MeroX- and XlinkX-based pipelines identify peptides of different length and charge state and with different ranking by score To better understand how XL-MS pipelines can be used to yield the most useful data, we investigated the similarities and differences of all crosslink identifications found in duplicate within each MeroX- and XlinkX-based pipeline. Firstly, we analysed all


10

Page 11 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


peptides by length, charge and score. Crosslinked peptides were mostly 5-30 amino acids long regardless of the program they were identified with (Figure 2A). The total length of each pair of crosslinked peptides found by MeroX, mostly between 10 and 25 amino acids, was however shorter than that found by XlinkX, which was mostly between 15 to 40 amino acids (Figure 2B). Despite that, there was little difference shown by MeroX or XlinkX in their capacities to identify crosslinked peptides where the peptides in each pair were of the same or different lengths (Figure 2C). The discrepancy in total size of the crosslinked species identified by MeroX and XlinkX was reflected in a difference in precursor charge. Figure 2D shows that the distribution of charge states for MeroX identifications was skewed toward charge states 3+ and 4+, whereas the majority of XlinkX identifications were 4+ up to 8+. Our observations are consistent with other reports, which indicated that MeroX preferentially identifies lower charge states 17 whereas XlinkX, from our examination of crosslinks reported from E. coli and HeLa lysates16, showed an average charge state of 4+. Whilst the higher charge states of crosslinked peptides found by XlinkX might reflect their total length, identifications will be influenced by software scoring systems. XlinkX uses a probabilistic scoring metric 11, 16, whereas MeroX includes probabilistic measures and features of crosslink spectra, such as doublet peaks and quality of fragment peaks 17. To gain insight into how MeroX and XlinkX score and rank crosslinked peptides from identical input data, we finally ranked and plotted 20 identifications common to MeroX and XlinkX from highest to lowest score. Interestingly, Figure 2E shows that the rank of the 20 identical crosslinked peptides (same peptide sequence, and link site) had no correlation when identified by MeroX or XlinkX. To summarise all the above, whilst both pipelines can identify crosslinked


11


Page 12 of 27

peptides, they show preferences in terms of total peptide size and charge and have differences in scoring systems that may influence identifications.

Figure 2: Merox and XlinkX show biases in identification of crosslinked peptides. A) Comparison of length of peptide 1 versus peptide 2, for all pairs of crosslinked peptides. Peptide 1 was fixed as the longer of the two peptides. B) Frequency distribution of the total length of peptides 1 and 2, for each crosslinked peptide pair. C) Distribution of percentage difference in length between peptide 1 and 2, relative to the total length of the two crosslinked peptides. D) Distribution of precursor charge states for crosslinked peptides identified by MeroX or XlinkX. E) Scatterplot of the rank order of 20 identical crosslinked peptides, found by both XlinkX and MeroX.

Inter- and intraprotein crosslinks in Hmt1p and Npl3p involving serine residues are biologically relevant MeroX can search for crosslinks involving lysine, serine, threonine and/or tyrosine residues whereas XlinkX searches only for crosslinks involving lysines. This resulted in notable differences between the sets of crosslinked peptides reported by both


12

Page 13 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


programs. Interestingly, of all 70 non-redundant crosslinks detected in and between Hmt1p and Npl3p, 22 were K-K crosslinks; these were all intraprotein. All interprotein crosslinks and some intraprotein crosslinks involved one or two serines, and to a much lesser extent, threonines or tyrosines (Supplementary Information Tables S-2 and S-3). It was important to assess the non-K-K crosslinks for their quality and to check their biological relevance. We first manually inspected all spectra of non-K-K crosslinks (Supplementary Information, Figure S-6). Almost all contained accurate parent ion masses, fragments consistent with their likely peptide sequences and the presence of two or more doublet peaks arising from crosslinker cleavage. There is strong evidence to suggest that these non-K-K crosslinks reflect relevant aspects of interactions and structure. Many non-K-K crosslinks were found between Hmt1p and the C-terminal SRGG-containing domain of Npl3p. Hmt1p methylates that region at up to 17 positions 24 and, given there are no lysines in the latter ~200 amino acids of Npl3, the crosslinks involving serines at different SRGG motifs likely reflect Hmt1p-Npl3p interactions during catalysis. The Npl3p intraprotein crosslinks in the SRGG region likely reflect homo-dimerization, given that (as noted previously) the SRGG region is essential for self-association. Hmt1p non-K-K intraprotein crosslinks were within the 25 Angstrom distance constraint of the linker, when C-alpha to C-alpha distances were measured on the structure (Table S-2). Together, the above suggests that the non-K-K crosslinks reported by MeroX can provide useful biological insights.

Analysis of lysine-crosslinked peptides by primary and secondary variables Having observed some biases in the characteristics of crosslinked peptides identified by MeroX and XlinkX, we sought to understand which upstream variables of XL-MS


13


Page 14 of 27

(choice of crosslinker, instrument and fragmentation type, and data analysis approaches) could contribute to these differences. Given that these variables are always interdependent, we studied each variable of an XL-MS experiment in context of the others. To do this, lysine-crosslinked peptides found as two or more exact replicates were first collated from all analyses. The variables associated with each crosslinked peptide were noted (being DSSO or DSBU, stepped-HCD on the QExactive plus or CID+ETD on the Tribrid Fusion Lumos, MeroX or XlinkX, Precursor or Reporter-ion search algorithm). To allow a series of in-depth comparisons, the total list of identifications was split according to a primary variable (e.g. MeroX or XlinkX) and the resulting groups investigated to understand whether any secondary variable (e.g. DSSO or DSBU crosslinker) were positively or negatively associated with identifications in each group. Below, we first undertake a detailed analysis of all lysine-crosslinked peptides to understand whether identifications made by data analysis platform (MeroX or XlinkX) were affected by choice of crosslinker, instrument and fragmentation, or search algorithm. We then analyse the lysine-crosslinked peptides with the primary variable of crosslinker type, then instruments and fragmentation strategy and finally search algorithm (Precursor or Reporter-ion).

Identifications unique to MeroX or XlinkX show small biases in crosslinker and Precursor vs Reporter-ion search All lysine-crosslinked peptide identifications were divided into two groups, using MeroX or XlinkX as a primary variable. These groups were analysed in two ways. First, replicate identifications were removed from each group to generate a non-

redundant set; note this set includes crosslinked peptides of identical sequences but


14

Page 15 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


of different crosslink sites. From this we determined which crosslinked peptides were found by both MeroX and XlinkX (Figure 3A). We also established which crosslinked peptides were unique to just one program. Second, we wished to understand whether crosslinked peptides were affected by secondary variables. To understand this, the full redundant set of crosslinked peptides identified was collated and we established which were found by both MeroX and XlinkX (Figure 3B). We then investigated the crosslinked peptides found by both programs with respect to the secondary variables of DSSO or DSBU, instrument and fragmentation by SteppedHCD or CID+ETD, Precursor or Reporter-ion search (Figure 3C). We finally investigated the crosslinked peptides from the full redundant set that were unique to MeroX or XlinkX with respect to the secondary variables (Figure 3D). A detailed description of how non-redundant and redundant sets of identifications were generated and compared is given in Figure S-5.

From the analytical steps above, the non-redundant set of 81 crosslinked peptides from Hmt1p and Npl3p had 48 (59%) found by both MeroX and XlinkX (Figure 3A). We also found 25 identifications (31%) that were unique to MeroX and 8 (10%)

unique to XlinkX (Figure 3A). It is clear, that for the two proteins studied here, use of two programs improved the yield of non-redundant lysine-crosslinked peptides.

In our experiment, any crosslinked peptide can be identified under a number of different circumstances (as termed here, secondary variables). To understand the importance of these variables, we analysed the redundant set of 296 crosslinked peptides (Figure 3B). A total of 209 (71%) crosslinked peptides were found by both MeroX and XlinkX; these were plotted with respect to secondary variables (Figure


15


Page 16 of 27

3C). MeroX identified each peptide more frequently, for every variable, suggesting it is more flexible than XlinkX. However, the 209 crosslinked peptides found by both MeroX and XlinkX showed little bias from secondary variables identifying proportionally similar crosslinks using DSSO, SteppedHCD and Precursor algorithm. Together, the results above suggest the existence of a core set Hmt1p and Npl3p crosslinked peptides; these can be identified by both MeroX and XlinkX under a variety of conditions.

There were a large number of crosslinked peptides in the redundant set that were

unique to Merox (67, 23%) and a smaller number that were unique to XlinkX (20, 6.8%) (Figure 3B). We investigated these to determine if their uniqueness was due to bias in the secondary variables of crosslinker, instrument and fragmentation, or of algorithm (Figure 3D). Both MeroX and XlinkX identified more crosslinks with DSSO than DSBU. However, MeroX depicted a small bias for DSBU; this has been previously noted by Arlt et al. (2015) to arise from MeroX favoring DSBU spectra and XlinkX favoring DSSO fragmentation spectra 20. For fragmentation itself, there was no bias in how SteppedHCD or CID+HCD affected identification in MeroX or XlinkX. Regarding Precursor or Reporter-ion search strategies, crosslinked peptides

unique to XlinkX were found equally with Precursor or Reporter-ion strategies, whereas 70% of those unique to MeroX were detected by a Precursor search strategy. Together, these results suggest that there are some minor differences in how MeroX and XlinkX algorithms can identify certain Hmt1p and Npl3p peptides.


16

Page 17 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3: Factors that contribute to the similarities and differences in identification between MeroX and XlinkX. A) Comparison of non-redundant, crosslinked peptides identified by MeroX and/or XlinkX. Peptides identified by both (overlap in Venn diagram) are of identical sequence and link site. B) Comparison of redundant, crosslinked peptides identified by MeroX and/or XlinkX. C) Redundant crosslinked peptides identified by both Merox and XlinkX (overlap in Venn diagram, panel B), and the secondary variables used for their analysis. The radar plot shows the number of identifications of which arose from each specific secondary variable. D) Redundant crosslinked peptides that are unique to either MeroX or XlinkX (those that do not overlap in Venn diagram, panel B), and the secondary variables used for their analysis. Data normalized against the number of identifications for each search engine, for each secondary variable, and represented as percent.

Effect of crosslinker, instrument and fragmentation, and the search algorithm on identification of crosslinked peptides


17


Page 18 of 27

Having investigated the differences between crosslinks generated by MeroX and/or XlinkX, we then investigated the degree to which choice of crosslinker, instrument and fragmentation or algorithm type affected outcomes. To facilitate this, lysinecrosslinked peptides found in two or more exact replicates were analysed in three ways, according to a primary variable. The first analysis grouped identifications by crosslinker type (DSBU or DSSO), the second by instrument and fragmentation type (CID+ETD or SteppedHCD) and the third by algorithm type (Reporter-ion or Precursor). Each primary variable was then investigated by secondary variables.

DSBU or DSSO crosslinkers have largely been used with MeroX and XlinkX search engines, respectively. Previous literature has suggested DSBU performs best under MS2 conditions 18, whereas DSSO achieves the greatest number of identifications in XlinkX using MS3 conditions16. Accordingly, we investigated if lysine-crosslinked peptides with DSBU or DSSO are more amenable to identification with different fragmentation or programs. A binary-tree analysis of all data (Figure S-4) highlighted that, for both DSBU and DSSO, that the number of identifications obtained with CID+ETD was greater than that obtained by SteppedHCD and that the number of identifications obtained for MeroX was greater than that for XlinkX. In a detailed analysis as per Figure 3, a total of 48 non-redundant crosslinks (60%) were identified from both DSBU and DSSO, whereas 7 and 26 crosslinks (total 40%) were unique to DSBU or DSSO respectively (Figure 4A). The high number of unique identifications, versus those identified by both crosslinkers, shows that it was valuable to use both DSBU and DSSO to maximise insight into Hmt1p and Npl3p. We then investigated whether crosslinked peptides identified by both DSBU and DSSO were affected by secondary variables. Interestingly, this showed that 71% of redundant crosslinked


18

Page 19 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


peptides were identified equally well with both DSBU and DSSO (Figure 4B) and that of these, there was little bias associated with any secondary variable (Figure 4C) although a greater number of crosslinks were identified with CID+ETD than SteppedHCD. However, for the crosslinks that were unique to DSBU or DSSO, there were effects present from the secondary variable of program. Specifically, 75% of crosslinked peptides unique to DSBU were from MeroX (Figure 4D) whereas for identifications unique to DSSO, ~60% of those arose from MeroX. However, there was little or no bias present from fragmentation type or search algorithm in the unique peptide crosslinks. To summarise, factors that affect identification outcomes with crosslinker DSBU or DSSO are the choice of analysis program (MeroX or XlinkX) and to a lesser extent, fragmentation.

Multiple fragmentation strategies can be used in XL-MS 20, 33. Here, we investigated whether identifications arising from the primary variable of fragmentation by SteppedHCD on a Q Exactive Plus or CID+ETD fragmentation on a Tribrid Fusion Lumos were affected by the secondary variables of crosslinker, data analysis program or search algorithm. Analysis of the non-redundant set of crosslinks, grouped by SteppedHCD or CID+ETD (Figure 4E), showed that there were 66 identifications of Hmt1p and Npl3p crosslinked peptides (74%) found by both SteppedHCD and CID+ETD. Yet 10 and 13 crosslinks (total 26%) were unique to CID+ETD or SteppedHCD, respectively. The combination of both fragmentation approaches was thus useful to maximize the non-redundant crosslinks detected for Hmt1p and Npl3p.

To understand the impact of secondary variables on identifications grouped by instrument and fragmentation, the redundant set of crosslinked peptides was


19


Page 20 of 27

analysed. Those identified by both SteppedHCD and CID+ETD (247 of 296, 83%) were not appreciably biased by the secondary parameters of crosslinker or program (Figure 4F, G, Tables S-2, S-3). However, the choice of algorithm influenced identification success in that SteppedHCD showed a strong bias towards Precursor ion whereas CID+ETD showed a bias in identification with Reporter-ion algorithm (Figure 4G). A greater number of identifications was also made by CID+ETD as compared to SteppedHCD. For identifications from the redundant set that were unique to SteppedHCD or CID+ETD, there was little bias from the secondary parameter of crosslinker (Figure 4H). However, and similar but more extreme than the redundant identifications for both SteppedHCD and CID/ETD, unique identifications by SteppedHCD were almost entirely from Precursor ion searches and this was different to unique CID+ETD identifications that were just ~50% from Precursor ion searches (Figure 4H). Manual inspection of the spectra associated with unique crosslinked peptides from SteppedHCD determined that 63% did not have doublets, and 46% had poor fragmentation (data not shown), likely explaining why they were not identified with a Reporter-ion approach. There is thus a clear association between choice of fragmentation and the best algorithm to use for analysis.

MeroX and XlinkX have algorithms that identify crosslinked peptides in Precursor ion mode17 but can also take advantage of the Reporter-ion peaks generated from MS-cleavable crosslinkers. To understand if Precursor and Reporter-ion algorithms were of equivalent utility, we grouped identifications by the primary variable of algorithm. Analysis of the non-redundant set of crosslinks, showed that 83% of identifications were identified by both algorithms (Figure 4I). Some 14.6% were


20

Page 21 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


unique to the Precursor algorithms, but only 2% unique to Reporter-ion algorithms. In the redundant set of crosslinked peptides, 93% were found by both Precursor and Reporter-Ion search, 6.8% were unique to the Precursor search and 0.7% unique to Reporter-Ion search. Crosslinks found by Reporter-ion algorithms were almost entirely a subset of those found by Precursor ion search. Examination of the effect of secondary variables, on the redundant set of peptides found by both algorithms (Figure 4 J,K), showed that Precursor algorithms generated markedly more crosslink identifications with SteppedHCD fragmentation. Other secondary variables did not appear to clearly associate with identification algorithm. For the very small number of crosslinked peptides that were unique to Precursor or Reporter-ion identifications (Figure 4 J, 20 and 2 crosslinked peptides respectively), the associations between SteppedHCD or CID+ETD were dramatic, with ~80% of unique identifications from Precursor searches arising from SteppedHCD but 100% of unique identifications from Reporter-ion searches arising from CID+ETD (Figure 4 L). Interestingly, the small number of crosslinked peptides that were unique to Precursor or Reporter-ion identifications were identified by MeroX but not by XlinkX. To better understand the predominance and presence/absence of reporter-ion peaks, all spectra were manually inspected. Identifications common to both algorithms had reporter-ion peaks within their spectra whereas only 40% of the identifications that were unique to the Precursor algorithm results had reporter-ions present (data not shown). Interestingly all XlinkX identifications, regardless of algorithm, contained at least one set of doublet reporter ions (Figure S-8). To summarise, Precursor and Reporter-ion algorithms generated largely overlapping sets of lysine-crosslinked peptides but ~15% more were generated by the Precursor algorithms. Further to this, the success


21


Page 22 of 27

of algorithm was not affected by crosslinker used although Precursor searches were more effective with SteppedHCD data.

Figure 4: Factors that contribute to the similarities and differences in crosslinked peptide identifications: crosslinker, fragmentation and algorithm. Panels A-D compare results between crosslinkers (DSBU and DSSO), panels E-H compare those between fragmentation (SteppedHCD on Q-Exactive and CID+ETD on Fusion Lumoc) and panels I-L compare those between algorithm type (Precursor and Reporter Ion). Relationships between data in each vertical group (A to D, E to H, I to L) are as per Figure 3.

Conclusions In this study we used XL-MS to investigate the enzyme-substrate interaction between methyltransferase Hmt1p and its substrate Npl3p. By using multiple XL-MS approaches in our experiment, namely two crosslinkers, two fragmentation


22

Page 23 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


approaches on two instruments, two programs and two algorithms therein, we were able to maximise insights into Hmt1p, Npl3p and their interaction. We were also able to investigate how the variables of XL-MS can influence the identification success of crosslinked peptides.

Hmt1p methylates Npl3p at 17 arginines, which are found in a disordered region containing tandem repeats of an SRGG motif 24, 30. The approach used here has successfully captured the transient enzyme-substrate interaction that occurs during methylation. We believe this is the first such observation of this process. We have found evidence of interaction between Hmt1p and multiple regions of the SRGG tandem repeat, and showed that this is where methylation of Npl3p takes place, also capturing an interaction that involves a disordered domain. We generated evidence of the multimerization of Hmt1p and of Npl3p, in agreement with the literature and the 3-D structure of Hmt1p. A single region, involving K246 of Hmt1p, was found as an interaction hotspot with multiple tandem repeats of Npl3p; this interaction is likely important for the process of methylation.

The use of MS-cleavable linkers in XL-MS, and the new software for identification of crosslinked peptides, is an emerging technique. Our multi-crosslinker, multiinstrument fragmentation and multi-software approach provided a more comprehensive set of crosslinked peptides than any single XL-MS approach. Our analysis has also, however, revealed that some combinations of experimental variables will likely create greater yield of crosslinks than others. A set of observations, arising from our analyses, is summarized below.


23


Page 24 of 27

1. DSBU and DSSO crosslinkers have a capacity to generate different K-K crosslinked peptides, but their identification relies on use of both MeroX and XlinkX. 2. The use of MeroX and XlinkX will increase coverage of K-K crosslinks present and MeroX can be used to identify useful non-K-K crosslinks. 3. Crosslinked peptides found by Reporter-ion algorithms are a large subset of those found by Precursor ion algorithms; however, some unique peptides can be found by Reporter-ion algorithms. 4. Choice of instrument and fragmentation technique must be appropriately matched with algorithm. SteppedHCD requires the use of Precursor ion algorithms whereas CID+ETD data will benefit from being analysed with either algorithm type.

Acknowledgements We thank Andrea Sinz, Philip Lössl, Fan Liu and Albert Heck and for technical assistance and for commentary on this manuscript. DLS was the recipient of an Australian Government Research Training Program scholarship. GH-S and MRW were recipients of funding from the Australian Research Council.

References 1. Benschop, J. J.; Brabers, N.; van Leenen, D.; Bakker, L. V.; van Deutekom, H. W.; van Berkum, N. L.; Apweiler, E.; Lijnzaad, P.; Holstege, F. C.; Kemmeren, P., A consensus of core protein complex compositions for Saccharomyces cerevisiae. Mol Cell 2010, 38 (6), 916-28. 2. Hein, M. Y.; Hubner, N. C.; Poser, I.; Cox, J.; Nagaraj, N.; Toyoda, Y.; Gak, I. A.; Weisswange, I.; Mansfeld, J.; Buchholz, F.; Hyman, A. A.; Mann, M., A human interactome in three quantitative dimensions organized by stoichiometries and abundances. Cell 2015, 163 (3), 712-23.


24

Page 25 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


3. Uetz, P.; Giot, L.; Cagney, G.; Mansfield, T. A.; Judson, R. S.; Knight, J. R.; Lockshon, D.; Narayan, V.; Srinivasan, M.; Pochart, P.; Qureshi-Emili, A.; Li, Y.; Godwin, B.; Conover, D.; Kalbfleisch, T.; Vijayadamodar, G.; Yang, M.; Johnston, M.; Fields, S.; Rothberg, J. M., A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature 2000, 403 (6770), 623-7. 4. Rappsilber, J.; Siniossoglou, S.; Hurt, E. C.; Mann, M., A generic strategy to analyze the spatial organization of multi-protein complexes by cross-linking and mass spectrometry. Anal Chem 2000, 72 (2), 267-75. 5. Kao, A.; Chiu, C. L.; Vellucci, D.; Yang, Y.; Patel, V. R.; Guan, S.; Randall, A.; Baldi, P.; Rychnovsky, S. D.; Huang, L., Development of a novel cross-linking strategy for fast and accurate identification of cross-linked peptides of protein complexes. Mol Cell Proteomics 2011, 10 (1), M110 002212. 6. Sinz, A.; Wang, K., Mapping protein interfaces with a fluorogenic crosslinker and mass spectrometry: application to nebulin-calmodulin complexes. Biochemistry 2001, 40 (26), 7903-13. 7. Dihazi, G. H.; Sinz, A., Mapping low-resolution three-dimensional protein structures using chemical cross-linking and Fourier transform ion-cyclotron resonance mass spectrometry. Rapid Commun Mass Spec 2003, 17 (17), 2005-14. 8. Schulz, D. M.; Ihling, C.; Clore, G. M.; Sinz, A., Mapping the topology and determination of a low-resolution three-dimensional structure of the calmodulinmelittin complex by chemical cross-linking and high-resolution FTICRMS: direct demonstration of multiple binding modes. Biochemistry 2004, 43 (16), 4703-15. 9. Kalkhof, S.; Ihling, C.; Mechtler, K.; Sinz, A., Chemical cross-linking and high-performance Fourier transform ion cyclotron resonance mass spectrometry for protein interaction analysis: application to a calmodulin/target peptide complex. Anal Chem 2005, 77 (2), 495-503. 10. Ihling, C.; Schmidt, A.; Kalkhof, S.; Schulz, D. M.; Stingl, C.; Mechtler, K.; Haack, M.; Beck-Sickinger, A. G.; Cooper, D. M.; Sinz, A., Isotope-labeled crosslinkers and Fourier transform ion cyclotron resonance mass spectrometry for structural analysis of a protein/peptide complex. J Am Soc Mass Spectr 2006, 17 (8), 1100-13. 11. Liu, F.; Rijkers, D. T.; Post, H.; Heck, A. J., Proteome-wide profiling of protein assemblies by cross-linking mass spectrometry. Nat Methods 2015, 12 (12), 1179-84. 12. Robinson, P. J.; Trnka, M. J.; Bushnell, D. A.; Davis, R. E.; Mattei, P. J.; Burlingame, A. L.; Kornberg, R. D., Structure of a Complete Mediator-RNA Polymerase II Pre-Initiation Complex. Cell 2016, 166 (6), 1411-1422 e16. 13. Kastritis, P. L.; O'Reilly, F. J.; Bock, T.; Li, Y.; Rogon, M. Z.; Buczak, K.; Romanov, N.; Betts, M. J.; Bui, K. H.; Hagen, W. J.; Hennrich, M. L.; Mackmull, M. T.; Rappsilber, J.; Russell, R. B.; Bork, P.; Beck, M.; Gavin, A. C., Capturing protein communities by structural proteomics in a thermophilic eukaryote. Mol Syst Biol 2017, 13 (7), 936. 14. Muller, M. Q.; Dreiocker, F.; Ihling, C. H.; Schafer, M.; Sinz, A., Cleavable cross-linker for protein structure analysis: reliable identification of cross-linking products by tandem MS. Anal Chem 2010, 82 (16), 6958-68. 15. Tang, X.; Bruce, J. E., A new cross-linking strategy: protein interaction reporter (PIR) technology for protein-protein interaction studies. Mol Biosyst 2010, 6 (6), 939-47.


25


Page 26 of 27

16. Liu, F.; Lossl, P.; Scheltema, R.; Viner, R.; Heck, A. J. R., Optimized fragmentation schemes and data analysis strategies for proteome-wide cross-link identification. Nat Commun 2017, 8, 15473. 17. Gotze, M.; Pettelkau, J.; Fritzsche, R.; Ihling, C. H.; Schafer, M.; Sinz, A., Automated assignment of MS/MS cleavable cross-links in protein 3D-structure analysis. J Am Soc Mass Spectr 2015, 26 (1), 83-97. 18. Weisbrod, C. R.; Chavez, J. D.; Eng, J. K.; Yang, L.; Zheng, C.; Bruce, J. E., In vivo protein interaction network identified with a novel real-time cross-linked peptide identification strategy. J Proteome Res 2013, 12 (4), 1569-79. 19. Arlt, C.; Flegler, V.; Ihling, C. H.; Schafer, M.; Thondorf, I.; Sinz, A., An Integrated Mass Spectrometry Based Approach to Probe the Structure of the FullLength Wild-Type Tetrameric p53 Tumor Suppressor. Angew Chem 2017, 56 (1), 275-279. 20. Arlt, C.; Gotze, M.; Ihling, C. H.; Hage, C.; Schafer, M.; Sinz, A., Integrated Workflow for Structural Proteomics Studies Based on Cross-Linking/Mass Spectrometry with an MS/MS Cleavable Cross-Linker. Anal Chem 2016, 88 (16), 7930-7. 21. Arlt, C.; Ihling, C. H.; Sinz, A., Structure of full-length p53 tumor suppressor probed by chemical cross-linking and mass spectrometry. Proteomics 2015, 15 (16), 2746-55. 22. Erce, M. A.; Low, J. K.; Hart-Smith, G.; Wilkins, M. R., A conditional twohybrid (C2H) system for the detection of protein-protein interactions that are mediated by post-translational modification. Proteomics 2013, 13 (7), 1059-64. 23. Hart-Smith, G.; Raftery, M. J., Detection and characterization of low abundance glycopeptides via higher-energy C-trap dissociation and orbitrap mass analysis. J Am Soc Mass Spectr 2012, 23 (1), 124-40. 24. Hart-Smith, G.; Low, J. K.; Erce, M. A.; Wilkins, M. R., Enhanced methylarginine characterization by post-translational modification-specific targeted data acquisition and electron-transfer dissociation mass spectrometry. J Am Soc Mass Spectr 2012, 23 (8), 1376-89. 25. Weiss, V. H.; McBride, A. E.; Soriano, M. A.; Filman, D. J.; Silver, P. A.; Hogle, J. M., The structure and oligomerization of the yeast arginine methyltransferase, Hmt1. Nat Struct Biol 2000, 7 (12), 1165-71. 26. Tardiff, D. F.; Abruzzi, K. C.; Rosbash, M., Protein characterization of Saccharomyces cerevisiae RNA polymerase II after in vivo cross-linking. Proc Natl Acad Sci U S A 2007, 104 (50), 19948-53. 27. McBride, A. E.; Conboy, A. K.; Brown, S. P.; Ariyachet, C.; Rutledge, K. L., Specific sequences within arginine-glycine-rich domains affect mRNA-binding protein function. Nucleic Acids Res 2009, 37 (13), 4322-30. 28. Erce, M. A.; Abeygunawardena, D.; Low, J. K.; Hart-Smith, G.; Wilkins, M. R., Interactions affected by arginine methylation in the yeast protein-protein interaction network. Mol Cell Proteomics 2013, 12 (11), 3184-98. 29. Combe, C. W.; Fischer, L.; Rappsilber, J., xiNET: cross-link network maps with residue resolution. Mol Cell Proteomics 2015, 14 (4), 1137-47. 30. Yagoub, D.; Hart-Smith, G.; Moecking, J.; Erce, M. A.; Wilkins, M. R., Yeast proteins Gar1p, Nop1p, Npl3p, Nsr1p, and Rps2p are natively methylated and are substrates of the arginine methyltransferase Hmt1p. Proteomics 2015, 15 (18), 320918. 31. McBride, A. E.; Weiss, V. H.; Kim, H. K.; Hogle, J. M.; Silver, P. A., Analysis of the yeast arginine methyltransferase Hmt1p/Rmt1p and its in vivo


26

Page 27 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


function. Cofactor binding and substrate interactions. J Biol Chem 2000, 275 (5), 3128-36. 32. Liu, Q.; Remmelzwaal, S.; Heck, A. J. R.; Akhmanova, A.; Liu, F., Facilitating identification of minimal protein binding domains by cross-linking mass spectrometry. Sci Rep 2017, 7 (1), 13453. 33. Kolbowski, L.; Mendes, M. L.; Rappsilber, J., Optimizing the Parameters Governing the Fragmentation of Cross-Linked Peptides in a Tribrid Mass Spectrometer. Anal Chem 2017, 89 (10), 5311-5318.

For TOC Only


27

Characterisation of the interaction between arginine methyltransferase

Recommend Documents