View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Impact of and Residues on the Dissociation of Intermolecularly Crosslinked

Myles W. Gardner and Jennifer S. Brodbelt Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, Texas, USA

The dissociation of intermolecularly crosslinked peptides was evaluated for a series of peptides with proline or aspartic acid residues positioned adjacent to the crosslinking sites ( residues). The peptides were crosslinked with either disuccinimidyl suberate (DSS) or disuccinimidyl L-tartrate (DST), and the influence of proline and aspartic acid residues on the fragmentation patterns were investigated for precursor ions with and without a mobile proton. Collisionally activated dissociation (CAD) spectra of aspartic acid-containing crosslinked ions, doubly- charged with both protons sequestered, were dominated by cleavage C-terminal to the Asp residue, similar to that of unmodified peptides. The proline-containing crosslinked peptides exhibited a high degree of internal ion formation, with the resulting product ions having an N-terminal proline residue. Upon dissociation of the doubly-charged crosslinked peptides, twenty to fifty percent of the fragment ion abundance was accounted for by multiple cleavage products. Crosslinked peptides possessing a mobile proton yielded almost a full series of b- and y-type fragment ions, with only proline-directed fragments still observed at high abundances. Interest- ingly, the crosslinked peptides exhibited a tendency to dissociate at the amide bond C-terminal to the crosslinked lysine residue, relative to the N-terminal side. One could envision updating computer algorithms to include these crosslinker specific product ions—particularly for precursor ions with localized protons—that provide complementary and confirmatory infor- mation, to offer more confident identification of both the crosslinked peptides and the location of the crosslink, as well as affording predictive guidelines for interpretation of the product-ion spectra of crosslinked peptides. (J Am Soc Mass Spectrom 2008, 19, 344–357) © 2008 American Society for Mass Spectrometry

ith tremendous advances in proteomics in the tides must be identified from unmodified ones, typically past decade, the interest in understanding pro- in complex mixtures such as enzymatic digests of the Wtein- interactions at a molecular level protein or protein-complexes, in which there may only be has expanded as new methods for exploring macromole- a few crosslinks per protein among an abundance of cules have emerged [1–3]. Chemical crosslinking of pro- unmodified peptide fragments. Once distinguished and teins coupled with mass spectrometric analysis has be- identified, collisionally activated dissociation (CAD) ex- come a more widely used method for determining periments on these species are performed to sequence the protein-protein interactions in recent years [4–6]. Protein peptide or peptides. Structural assignment of the crosslinking is a low-resolution, structural analysis tech- crosslinked peptides is still not routine despite the emer- nique that allows one to determine distance constraints gence of several sequencing algorithms [17–22]. In fact, within single [7, 8], or protein-ligand complexes one of the main pitfalls of mass spectrometric strategies [9–13]. Crosslinked proteins are typically subjected to for examination of crosslinked peptides is the lack of enzymatic digestion to create more easily interpreted comprehensive rules for interpretation of the product ion peptide segments upon subsequent tandem mass spectro- spectra. Finally, with the known sequence tags, the site metric analysis. Crosslinking has also been utilized to and residue of chemical modification must be pinpointed identify protein folds [6], conformational changes of pro- to estimate distance constraints. teins upon activation [14, 15], and RNA-protein interac- Despite the enormous number of investigations that tions [16]. While the chemical crosslinking experiment is have elucidated the fragmentation pathways of conven- often rather straightforward, the technique is restricted by tional tryptic peptides, to date only a few studies have several limitations which must be overcome to obtain the focused on mapping the dissociation trends of crosslinked desired structural information. First, the crosslinked pep- peptides in a systematic fashion [20, 23, 24]. Initial work by Schilling et al. demonstrated that dissociation of intermo- lecularly crosslinked peptides typically resulted in amide Address reprint requests to Dr. J. S. Brodbelt, Department of Chemistry and Biochemistry, The University of Texas at Austin, 1 University Station bond cleavage of one peptide, with the crosslink and A5300, Austin, TX 78712, USA. E-mail: [email protected] second peptide acting as a modification [20]. In addition,

Published online November 19, 2007 © 2008 American Society for Mass Spectrometry. Published by Elsevier Inc. Received August 21, 2007 1044-0305/08/$32.00 Revised November 2, 2007 doi:10.1016/j.jasms.2007.11.006 Accepted November 6, 2007 J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 345 they observed product ions unique to crosslinked pep- Experimental tides, including two cleavage fragment ions such as b/b ions, in which two b-type fragment ions were linked Reagents together. A more recent study investigating the effects of The following five peptides were synthesized at the lysine specific crosslinkers and precursor ion charge ob- Protein Microanalysis Facility at the University of Texas served that without a mobile proton, many of the product at Austin and purified by reversed-phase liquid chro- ions formed were not of the a-, b-, or y-type [23]. Gaucher matography: Ac-AAAKAAAAR (AKA), Ac-AAD- et al. also noticed that intramolecularly crosslinked pep- KAAAAR (DKA), Ac-AAAKDAAAR (AKD), Ac-AAP- tides exhibited enhanced dissociation at the peptide amide KAAAAR (PKA), and Ac-AAAKPAAAR (AKP). The bonds adjacent to the crosslinked lysine residues [23]. peptides neurotensin (Pyr-LYENKPRRPYIL) and While many fundamental studies and statistical anal- ␣ -MSH (Ac-SYSMEHFRWGKPV-NH2) were purchased yses have examined the influence of content from Bachem (Torrance, CA). The crosslinker disuccin- on the gas-phase dissociation of unmodified peptides imidyl suberate (DSS) was purchased from Sigma- [25–29], no similar work to date has been performed for Aldrich (St. Louis, MO), and disuccinimidyl L-tartrate crosslinked peptides. It has been shown that certain (DST) from Toronto Research Chemicals (North York, amino acids, including aspartic acid and proline, en- Ontario, Canada). All other chemicals and solvents hance or promote specific backbone cleavage of pep- were obtained from Fisher Scientific (Fairlawn, NJ) tides [30–38]. The mobile proton model, which pro- except for extra dry dimethylsulfoxide (DMSO) (Sigma- poses that peptide fragmentation is driven by charge- Aldrich). directed cleavages, has shown that higher energies are required to dissociate peptide ions in which the protons are sequestered at basic sites, particularly Chemical Crosslinking residues [26, 30, 38–40]. For such peptides which do not possess a mobile proton and have aspartic acid resi- Stock solutions of each peptide were prepared at 5.0 mM dues, enhanced cleavage at the Asp-Xxx bond has been in 20.0 mM sodium phosphate buffer, pH 7.5, and the observed [31, 41, 42]. Work by the Wysocki group has crosslinkers were freshly prepared at 20.0 mM in DMSO demonstrated that the acidic side-chain of aspartic acid directly before use. The crosslinking reaction was per- formed by adding 8.0 ␮L of each peptide to 2.5 ␮Lof directs fragmentation, site-specifically, at the C-termi- ␣ ␤ nal amide bond [33, 43]. Similar dissociation trends crosslinker, for a molar ratio of 4:4:5 ( peptide: peptide: were also observed for containing pep- crosslinker). The reaction was allowed to proceed at room tides [33]. Proline residues enhance dissociation N- temperature for 1 h, after which the reaction mixture was terminally, at the Xxx-Pro amide bond, often yielding desalted by solid-phase extraction using 50 mg tC18 Wa- intense y-type ions, and has been observed for the ters (Milford, MA) Sep-Pak cartridges. dissociation of small, singly-charged peptides as well as multiply-charged proteins [32, 44]. This site-specific Methods and Instrumentation cleavage was first attributed to the high gas-phase basicity of the proline residue [45], but has also been Samples were diluted to a concentration of 20 ␮Min theorized to be due to unfavorable ring strain of a 70/30/1 MeOH/H2O/HOAc (vol/vol/vol) for ESI-MS proposed bicyclic structure of the b-type ion that would analysis. Experiments were performed on a Ther- be formed upon cleavage C-terminal to the proline [46]. moFinnigan LCQ Duo (San Jose, CA) quadrupole ion The effects of amino acid content on the gas-phase trap with the standard electrospray source. Solutions fragmentation of crosslinked peptides, in particular were infused into the mass spectrometer at 3 ␮L/min residues which are known to promote site-specific using a Harvard Apparatus PHD 2000 syringe pump dissociation along the peptide backbone, have not been (Holliston, MA). Ionization and ion optic conditions analyzed. Knowledge of the dissociation trends of were optimized once for each precursor charge state crosslinked peptides, especially with regards to specific and were not retuned for each individual crosslinked amino acids, would be useful for improving computer peptide pair. Precursor ions were activated for 30 ms at algorithms to aid in the interpretation of product ion the typical qz value of 0.25 for CAD and the collision spectra of these ions. In this work, we investigate the energy was adjusted such that precursor ion abundance influence of aspartic acid and proline on the dissocia- was less than 10% of the total ion abundance. Typical tion of intermolecularly crosslinked peptides. As un- CAD energies for doubly-charged crosslinked peptide modified peptides tend to exhibit enhanced fragmenta- ions ranged from 1.08 to 1.24 V (ϳ30% normalized tion C-terminally to aspartic acid residues for precursor collision energy), and 0.57 to 0.62 V (ϳ20% normalized ions with sequestered protons, and N-terminally to collision energy) for the triply-charged species. The , these residues were positioned adjacent to the precursor ion isolation window was set to 4 m/z units to modified by commonly used primary amine- retain isotopic profile to aid in determining the charge specific crosslinkers. The impact of these residues on state of the product ions. For some of the doubly- the directionality of cleavage relative to the crosslinked charged crosslinked peptides, the activation qz-value lysine is also reported. was decreased to 0.22 from 0.25 to effectively trap the 346 GARDNER AND BRODBELT J Am Soc Mass Spectrom 2008, 19, 344–357

low-mass 3bfragment ion; the dissociation efficiency and fragment ion abundances were not significantly affected by this change. All collisionally activated dis- sociation experiments were collected in triplicate.

Data Analysis Relative product ion abundances were calculated by dividing each product ion peak area by the total frag- ment ion abundance. Fragment ions were grouped into different product ion types: y-ions, b and a-ions, crosslink amide bond cleavage products, lysine immo- nium ions, internal ions, other double cleavage prod- ucts (b/b, b/y, and y/y), and specific crosslinker cleav- age products (e.g., dissociation of the C7–C8 bond of DST). Peaks were defined as having a minimum relative intensity of 1% and a minimum area of 0.1% of the total area in the spectrum. Water and ammonia losses from each product ion type were included in the correspond- Ϫ Scheme 1 ing base group (e.g., peak arean NHof 3 iona y were included in the y-ion group). Neutral losses including water, ammonia, and 2 fromCO the precursor ion wereResults and Discussion excluded from the total fragment ion abundance for simplicity. To determine the effect of Asp andThe dissociationPro of crosslinked peptides containing either residues in regards to cleavage at the site of a crosslink-proline or an aspartic acid residue was investigated to ing, K4, the directionality of cleavage bias to determinethe lysine if these species follow similar fragmentation residue was calculated for both the crosslinked peptidespathways as unmodified peptides. Poly- sequences and the unmodified species using a method previouslywere used as “filler” residues to minimize competing suggested by Tabb et[28] .al. The C-terminal cleavagefragmentation pathways. The peptides were crosslinked bias was determined by taking the ratio of thewith differ-either DSS or DST (1Scheme), two widely used ence in C-terminal and N-terminal fragment ions,crosslinkers rela- which react with primary amines such as free tive to the lysine residue, to the sum of theseN-termini fragment or the␧-amine of lysine residues. Since the ion peak areas. For example, the y-ion C-terminalN-termini bias of these peptides were acetylated, it can be was calculated as follows: assumed that the site of crosslinking is between K4 of each peptide. A proline or aspartic acid residue was positioned ϩ Ϫ ϩ on either the C- or N-terminal side of the crosslinked (Ay5␣ Ay5␤) (Ay6␣ Ay6␤) C ϭ (1) lysine to determine whether the crosslink influenced the y ϩ ϩ ϩ (Ay5␣ Ay5␤) (Ay6␣ Ay6␤) dissociation trends of the peptides containing either of these two residues. The resulting doubly-charged crosslinked peptides are assumed to have both protons where A is the peak area of the respective y-ions; the y i 6 sequestered on the C-terminal arginine residues[30], ions are products of cleavage N-terminal to the whereas the triply-charged species possess a mobile pro- crosslinked lysine, and theions y are a result of 5 ton. The crosslinked peptides in bothϩ andthe ϩ3 2 C-terminal cleavage for the nine-residue peptides ana- charge states were analyzed by CAD to compare the lyzed in this study. ␣ Theand ␤ subscripts refer to effects of the proline or aspartic acid residues in different cleavage of the␣ and ␤ peptides, respectively, and are charge states. The product ions formed were labeled explained in more detail in the results and discussion. A according to the nomenclature suggested in references positive value indicates a fragmentation bias C- [20] and [23]. In brief, the two intermolecularly crosslinked terminally to the lysine residue, whereas a negative peptides are labeled ␣ and ␤, referring to the first and value signifies an N-terminal bias. Similarly, for the second peptide listed, respectively. Cleavage along the b-ions, the cleavage bias was calculated using the backbone of the ␣ peptide yielding a b-ion is labeled as following formula: bn␣, and similar notation is used other fragment ion types, as well as for the ␤ peptide. Product ions of both peptides ϩ Ϫ ϩ (Ab4␣ Ab4␤) (Ab3␣ Ab3␤) crosslinked together are labeled with the fragmentation C ϭ (2) b nomenclature for each cleavage (e.g., b4␣y7␤). Lysine im- (A ␣ϩ A ␤) ϩ (A ␣ϩ A ␤) b4 b4 b3 b3 monium ions, which maintain the crosslink to the other ␣ peptide, are labeled as KL , where L represents the ␣ where the 4 ionsb are the C-terminal cleavage productscrosslinker molecules (e.g., DSS or DST) and is the other and the 3 ionsb are the N-terminal fragments. peptide. Due to the repeating alanine sequences of the J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 347

Figure 1. CAD spectra of DSS intermolecularly crosslinked peptides. Precursor ions were (a)[AKA ϩ ϩ ϩ ϩ AKA ϩ DSS ϩ 2H]2 ,(b)[AKA ϩ AKA ϩ DSS ϩ 3H]3 ,(c)[AKD ϩ AKD ϩ DSS ϩ 2H]2 ,(d) ϩ ϩ [AKD ϩ AKD ϩ DSS ϩ 3H]3 ,(e)[PKA ϩ PKA ϩ DSS ϩ 2H]2 , and (f)[PKA ϩ PKA ϩ DSS ϩ ϩ 3H]3 . An asterisk is used to signify the precursor ion. Several double cleavage product ions could not be uniquely identified due to isobars inherent to the peptide sequences and are labeled with one possible identity preceded by a double dagger symbol. peptides studied, certain fragment ions cannot be specifi- dances of singly-charged crosslinked b-ions increased cally identified as either cleavage of the ␣ peptide or the ␤ as the amide bond cleavage site moved towards K4, peptide as the products are isobars, and are labeled with with the exception of the b8␣ fragment ion, which was subscript ␣/␤ to represent such cases. observed as the most abundant b-ion. Without the mobile proton several double cleavage products, such Dissociation of DSS-Crosslinked Peptides as b4␣y8␤ and b5␣b7␤, were also formed at similar abundances to the single cleavage fragment ions. Representative CAD spectra of the homo-intermolecu- Crosslinked lysine immonium ions were detected after larly DSS-crosslinked peptides are shown in Figure 1. the loss of neutral ammonia, as had previously been For the control crosslinked peptide, [Ac-AAAKAAAAR observed by Gaucher et al. [23], and cleavage of the ϩ ϩ DSS ϩ Ac-AAAKAAAAR ϩ 2H]2 , a series of crosslink amide also occurred, yielding the L␣13 ion of singly-charged non-crosslinked y-ions were observed in m/z 980.5. Dissociation of the triply-charged species increasing abundance as the site of cleavage moves yielded a more easily interpretable product ion spec- Ͼ Ͼ closer to the crosslinked lysine residue (i.e., y5␣ y4␣ trum in which greater than 95% of the fragment ions y3␣)(Figure 1a). Similar to this y-ion series, the abun- were single cleavage b- and y-type ions (Figure 1b). A 348 GARDNER AND BRODBELT J Am Soc Mass Spectrom 2008, 19, 344–357

ϩ full series of y-ions and b-ions were observed,PAAAR both ϩ DSS ϩ Ac-AADKAAAAR ϩ 2H]2 . Cleavage ␤ modified (y6␣—y8␣ and 4b␣–b8␣) and noncrosslinked of the amide bond between D3 and peptideK4 of the 2ϩ (y1␣–y5␣ and 2b␣–b3␣), but the relative abundances (Ac-AADKAAAAR),of thus resulting in6 ␤ theion, ywas the these fragment ions did not follow any strictmost trend. abundant pathway and the two proline-directed 2ϩ Triply-charged crosslinked y-ions were observed fragmentsand were the next two most abundant5␣ and ions, y ϩ the crosslinked b-ions were also detected in theb 4␣.higher, Also present, but to a much lesser degree, were the 2ϩ, charge state compared with the dissociation otherof thecrosslinked b-ions as well as a lysine immonium analogous doubly-charged precursor ions. In general,fragment crosslinked to Ac-AADKAAAAR. It should be for both the b- and y-ions, as the cleavage notedsite movedthat due to the alanine sequences present in both further away from the crosslink, the fragmentpeptides ion which create isobaric species, certain fragment abundance increased. At a much lower abundance,ions thecould not be identified conclusively as to whether the ␣ KL -17 product ion was also detected. bond cleavage occurred within␣ orthe the ␤ peptide (e.g., CAD of the homo-intermolecularly DSS-crosslinkedb5␣ has the same mass as5␤ fragmentthe b and as such is Ac-AAAKDAAAR peptide yielded very few sequencelabeled as5 ␣/b␤ in Supplemental Figure S1a). Dissociation ions for the doubly-protonated precursor ion, as ofshown the triply-charged DSS crosslinked peptide resulted in in Figure c.1 With both protons sequestered onan thealmost complete series of b- and y-type fragment ions, C-terminal arginine residues, aspartic acid directedas evident in Supplemental Figure S1a. With the addition fragmentation dominates and the two productsof ofthe mobile proton, dissociation ␤ ofpeptide the C- cleavage C-terminal to D5 are observed at terminala high to the Asp residue did not dominate the spec- abundance—y4␣ and 5b␣. The other three crosslinkedtrum, and was in fact one of few sequence fragment ions b-ions were also detected, but at a much lowernot observed.abun- Proline-directed fragmentation still yielded 2ϩ 2ϩ dance. Only one other y-ion is observed in thetwo productof the most abundant product andions, y␣ . b This 4␣ 5 ion spectrum. Cleavage N-terminally to the asparticdata suggests that to obtain fragmentation information acid residue yielded the product5 ␣ andions 4b␣y, but at essential to locate the site of crosslinking as well as to a much lower abundance. The triply-charged precursorsequence the two linked peptides, dissociation of more ion possesses a mobile proton, and dissociation highly-chargedof this crosslinked peptides would be beneficial. species was not dominated by a single pathway; rather, a full series of b- and y-ions were seen limitedDissociation only by of DST-Crosslinked Peptides the low-mass cut-off inherent to resonant excitation in ion trapsFigure ( d).1 Unlike the CAD spectrum forThe the DST-crosslinked peptides exhibited similar disso- Ac-AAAKAAAAR crosslinked peptide, no distinctciation pathways as the DSS-linked species, as shown in trends were observed regarding fragment ion Figureabun- , 2 with one notable exception in that cleavage of dances. Upon CAD, the crosslinked [Ac-AAPKAAAARthe C7–C8 bond of the DST crosslinker was observed ϩ ϩ DSS ϩ Ac-AAPKAAAAR ϩ 2H]2 peptide yielded for the doubly-charged precursor ions (described later an abundance of double cleavage products as evidentin this in section). The doubly-charged DST homo-inter- Figure e.1 Cleavage of the amide bond N-terminalmolecularly to crosslinked Ac-AADKAAAAR peptide the proline residue, P3, dominated the dissociationpredominantly of dissociated via Asp-directed fragmenta- this ion producing crosslinked internal 7␣PKA),ions (y tion, C-terminal to D3, as Figureseen a,in2 yielding the L Ϫ 2ϩ lysine immonium ions linked to7␣ ionthe (Kyy7␣ y6␣ ion. At much lower abundances, double cleavage 17), and a seriesn ␣y7of␤ fragmentb ions. Most of theseproducts, including ones involving cleavage C-terminal fragment ions are redundant as they do not toprovide the aspartic residues of both peptides (y6␣y6␤) and additional information to identify the location crosslinkedof the b-ions were observed. CAD of the triply- crosslink nor sequence the constituent peptides. chargedDisso- species, presumed to have a mobile proton, ciation of the triply-charged species was similarlyyielded dom- a full series of b- and y-type fragment ions, inated by cleavage N-terminal to the proline againresidue only limited by the low-mass cut-off (Figure 2b). yielding the7 ␣ yfragment; in addition a high abundanceWith the mobile proton, the Asp-directed fragmenta- of the 8␣ yproduct was observed as shownFigure in f. 1 tion which dominated the CAD mass spectrum of the No cleavage between P3 and K4 was detecteddoubly-charged for the crosslinked peptide was circumvented. DSS crosslinked Ac-AAPKAAAAR, regardless of Fragmentpre- ions unique to crosslinked peptides, includ- cursor ion charge state. ing products from the dissociation of the crosslink The hetero-crosslinked peptides containing one prolineamide bond (L␣9 and L␣5), and crosslinked lysine ␣ ␣ and a single aspartic acid residue exhibited dissociationimmonium ions (KL and KL -17) were also detected. that followed both Asp- and Pro-directed pathways.For CAD all of the DST-crosslinked peptides analyzed, frag- mass spectra of these doubly-charged DSS-crosslinkedmentation of the C7–C8 bond of the DST crosslinker species were typically dominated by fragmentationoccurred, C- as illustrated in Scheme2. Dissociation of this terminal to the aspartic acid residue and to a lesserbond yieldsdegree, product ions corresponding to ϩ56 and N-terminal to the proline. A representative CAD spectrumϩ58 Da adducts of the unmodified peptides, and they is shown in Supplemental Figure 1a (which can arebe labeledfound as L␣7 and L␤7 ions. (It should be noted that in the electronic version of this article) for [Ac-AAAK-there is a 1 m/z unit difference between the precursor J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 349

Figure 2. CAD spectra of DST intermolecularly crosslinked peptides. Precursor ions were (a)[DKA ϩ ϩ ϩ ϩ DKA ϩ DST ϩ 2H]2 ,(b)[DKA ϩ DKA ϩ DST ϩ 3H]3 ,(c)[AKP ϩ AKP ϩ DST ϩ 2H]2 ,(d) ϩ ϩ [AKP ϩ AKP ϩ DST ϩ 3H]3 ,(e)[PKA ϩ AKD ϩ DST ϩ 2H]2 , and (f)[PKA ϩ AKD ϩ DST ϩ ϩ 3H]3 . The first peptide listed is referred to as the ␣ peptide and the second peptide as ␤; ␣/␤ nomenclature is used to signify that cleavage could have occurred on either peptide. Internal ions that could not be conclusively identified due to isobaric species are labeled with inverted triangle symbols. An asterisk is used to signify the precursor ion. ion and L␣7 fragment ions in Figure 2a–d.) For the Collisionally activated dissociation of [Ac-AAP- ϩ second proline-containing peptide analyzed in this KAAAAR ϩ DST ϩ Ac-AAAKDAAAR ϩ 2H]2 more study, Ac-AAAKPAAAR, dissociation of the doubly- clearly shows fragmentation of the C7–C8 bond of DST, charged DST-crosslinked peptide predominantly pro- producing the corresponding L␣7 and L␤7 ions, as duced three fragment ions, L␣7 and the two fragments shown in Figure 2e. Aspartic acid-directed fragmenta- ϩ corresponding to proline directed dissociation, y5␣ and tion dominated the product ion spectrum yielding y4␣/␤ ϩ b4␣ (Figure 2c). The triply-charged Ac-AAAKPAAAAR and b5␣/␤, and numerous proline-directed fragments ϩ DST ϩ Ac-AAAKPAAAR ion also showed preferen- were detected between m/z 600 and 1300, including tial cleavage between K4 and P5, while the other b- and internal ions of the form PKAn crosslinked to bn␤ ions y-type ions had abundances similar to that observed for (labeled with an inverted triangle symbol) and y7␣ the 2ϩ charged precursor (Figure 2d). Regardless of the containing fragments. The influences of both Asp and charge state, proline-directed fragmentation dominated Pro were observed by the formation of certain product 2ϩ the CAD spectrum of the Ac-AAAKPAAAR ions, including y7␣b5␤ , analogous to those observed for crosslinked peptides. the DSS-crosslinked species. The triply-charged DST 350 GARDNER AND BRODBELT J Am Soc Mass Spectrom 2008, 19, 344–357

crosslinked peptides, two types of fragment ions unique to crosslinked peptides, those involving cleavage of the crosslink amide bond and crosslinked lysine immo- nium ions, accounted for 2.5% and ϳ6% of the product ion abundance, respectively. Crosslinked peptides con- taining Ac-AAPKAAAAR yielded a high abundance of internal ions; ϳ11% of all the fragments observed were internal ions, compared to 2% for the non-Ac-AAP- KAAAAR crosslinked peptides. A majority of these internal ions had the proline residue at the N-terminus. Dissociation of the various triply-charged DSS-linked peptides, all of which possess a mobile proton, yielded a much higher abundance of single-cleavage a-, b-, and y-type fragment ions—on average greater than 95% relative abundance, as shown in Figure 3b. The total abundance of the crosslinked lysine immonium ions was less than 1%, and cleavage of the crosslink amide bond was observed at less than 0.1% abundance. For the triply-charged aspartic acid- or the proline-containing crosslinked peptides, there were no major differences in fragment ion abundances except that y-ions were ob- served at over 76% relative abundance for the Ac- AAPKAAAAR crosslinked species, due to preferential cleavage N-terminal to the proline residue yielding an Scheme 2 unusually high abundance of y7␣ fragments. No signif- icant differences in the multiple cleavage products were hetero-crosslinked peptide yielded a series of doubly- detected between the various crosslinked peptides. charged crosslinked b-ions, a series of crosslinked y- The DST-crosslinked peptides, particularly for spe- ions, and a series of uncrosslinked y-ions for both cies without a mobile proton as noted earlier, exhibited peptides, as evident in Figure 2f. The aspartic acid far more unique dissociation pathways compared to the effects were not observed when the precursor ion had a DSS-linked peptides. Fragmentation of the C7–C8 bond mobile proton, but fragment ions with N-terminal pro- of the DST crosslinker occurred for all of these lines were detected at high abundances, in particular crosslinked peptides in the 2ϩ charge state, accounting y7␣/␤. The DST specific cleavage of the C7–C8 bond was for over 21% of the total fragment ion abundance, as also observed, but at lower abundances compared to shown in Figure 4a, and these fragments do not provide that observed for the analogous doubly-charged precur- any diagnostic information for determining the sor ion. Dissociation of the amide bond N-terminal to crosslink location or sequencing the peptide. Similar to the crosslinked lysine for either constituent peptide, the DSS-crosslinked peptides, doubly-charged DST- Ac-AAPKAAAAR or Ac-AAAKDAAAR, to yield the linked species yielded on average 60% single cleavage expected y6␣ or y6␤ fragments did not occur, suggesting a-, b-, and y-ions, with the remainder accounted for by a cleavage bias due to the crosslink. multiple cleavage products and crosslinker specific product ions. The latter product ions, excluding ones Fragment Ion Type Abundances stemming from dissociation of the DST linker at the C7–C8 bond, were observed at ϳ15% relative abun- The relative abundance of each fragment ion type was dance, similar to that for the DSS-crosslinked species. calculated to shed light on the general trends of the Various internal ions were also observed at similar dissociation of these proline- and aspartic acid-contain- abundances for the DST-linked peptides than for the ing crosslinked peptides. As shown in Figure 3a for the DSS-crosslinked products, 3% relative abundance com- doubly-charged DSS-crosslinked peptides, a-, b-, and pared to ϳ5%, respectively. The DST-crosslinked prod- y-type fragment ions, on average, accounted for ϳ75% ucts containing the Ac-AAPKAAAAR peptide did ex- of the total product ion abundance. For aspartic acid- hibit a higher degree of internal ion formation than the containing crosslinked peptides, this abundance in- other peptides studied, but the difference was not as creased to over 85% as Asp-directed fragmentation of pronounced as that for the DSS-crosslinked products. these doubly-charged precursor ions, in which both Dissociation of the triply-charged DST-crosslinked pep- protons are sequestered, produced predominantly sin- tides, as summarized in Figure 4b, yielded similar gle cleavage b- and y-type ions. Multiple cleavage results as the analogous DSS-linked species. Across the fragments, including internal ions, b/b, b/y, y/y, and series, over 92% of the fragment ion abundance con- linked lysine immonium ions, accounted for 19% of all sisted of single bond cleavage a-, b-, and y-ions. Frag- the fragment ions. For the doubly-charged DSS- ment ions due to cleavage of the DST crosslinker at the J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 351

Figure 3. Relative fragment ion abundance for the DSS crosslinked peptides in the (a)2ϩ charge state and (b)3ϩ charge state. Each category represents the abundance of that product ion type relative to the total fragment ion abundance minus neutral losses from the precursor ion. Errors are less than 0.5% relative abundance.

C7–C8 bond totaled less than 1% relative abundance, The addition of the mobile proton to these DSS- suggesting that this fragmentation pathway is preferred and DST-crosslinked peptides increased the abun- only for precursor ions without a mobile proton. The dance of useful sequence ions, ones which are easily only modest difference between the fragmentation pat- interpretable. Also, aspartic acid-directed fragmenta- terns of the triply-charged DSS- and DST-linked pep- tion of the triply-charged ions did not occur, thus tides was an increased abundance of double cleavage yielding a wider array of single cleavage products. products and crosslinked lysine immonium ions, from Proline-directed fragmentation still dominated the 3% to 6%. CAD spectra of the triply-charged crosslinked pep- 352 GARDNER AND BRODBELT J Am Soc Mass Spectrom 2008, 19, 344–357

Figure 4. Relative fragment ion abundance for the DST crosslinked peptides in the (a)2ϩ charge state and (b)3ϩ charge state. Each category represents the abundance of that product ion type relative to the total fragment ion abundance minus neutral losses from the precursor ion. Errors are less than 0.5% relative abundance. tides, however, a full series of b- and y-ions were still Effect of Crosslinkers on Cleavage Bias Adjacent to observed, with the exception of cleavage C-terminal the Modified Lysine to proline residues. The degree of multiple-cleavage products, particularly the abundance of internal ions As evident from the dissociation of the intermolecularly for the proline containing peptides, was greatly re- crosslinked peptides, the aspartic acid or proline residue duced for dissociation of the more highly-charged tends to direct peptide fragmentation. One earlier study crosslinked peptide ions. indicated a qualitative enhancement of cleavage of the J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 353

Figure 5. The C-terminal cleavage bias relative to the lysine residue, for both the unmodified peptides in the 1ϩ and 2ϩ charge state and the homo-intermolecularly crosslinked peptides in the 2ϩ and 3ϩ charge state, for (a) y-ions and (b) b-ions. amide bond adjacent to the crosslinked lysine for intramo- unmodified peptides, 2ϩ charge state for crosslinked lecularly crosslinked peptides, so we aimed to investigate peptides), the influences of proline and aspartic acid on a possible preference for intermolecularly crosslinked spe- y-ion formation were not greatly affected upon crosslink- cies [23]. The directionality bias of cleavage to the ing (Figure 5a). The y-ion C-terminal bias was only en- crosslinked lysine was determined for both the unmodi- hanced for the Ac-AAAKPAAAR and Ac-AAAKAAAAR fied peptides (1ϩ and 2ϩ) and the homo-intermolecularly peptides after crosslinking. All of the doubly-charged crosslinked peptides (2ϩ and 3ϩ) as shown in Figure 5.A crosslinked peptides preferentially dissociated via cleav- positive C-terminal cleavage bias value indicates that age of the C-terminal amide bond to the lysine residue, more fragment ions were formed from cleavage of the except for the homo-intermolecularly crosslinked Ac- amide bond C-terminal to the crosslinked lysine, whereas AADKAAAAR peptide; these species fragmented via a negative value indicates a bias towards forming product Asp-direction dissociation C-terminal to D3, or N-termi- ions from dissociation of the N-terminal amide bond. For nally to the modified lysine residue. With the addition of species without a mobile proton (e.g., 1ϩ charge state for a mobile proton, the Asp-containing crosslinked peptides 354 GARDNER AND BRODBELT J Am Soc Mass Spectrom 2008, 19, 344–357 did not exhibit any significant cleavage bias towards the presence of at least one C-terminal arginine residue. The formation of y-ions. The crosslinked peptides with proline location of the may explain the greater C- residues still displayed signs of a C-terminal fragmenta- terminal cleavage bias observed for the b-ions compared tion bias in the 3ϩ charge state; however, these results can to y-ion formation, particularly for the triply-charged be explained by preferential cleavage N-terminally to precursor ions without a proline residue. However, in proline residues. With the proline residue positioned most protein crosslinking experiments, crosslinked pep- N-terminally to the lysine (e.g., Ac-AAPKAAAAR), cleav- tides with C-terminal arginines will be produced after age of the bond between P3 and K4 is suppressed, thus enzymatic digestions with trypsin, and the trends ob- decreasing the y6␣ product ion abundance relative to y5␣ served in this study should be considered in sequence fragment ions. For the other proline containing peptide, identification software.

Ac-AAAKPAAAR, formation of the y5␣ ion is enhanced For the unmodified peptides, the mobile proton is relative to y6␣ due to Pro-directed dissociation between K4 assumed to be located on the second most basic site, the and P5. ␧-amine group of the lysine residue. This proton, when Analysis of the b-ions detected showed that the un- mobilized upon ion activation, could easily promote modified peptides all preferentially dissociated N-termi- charge-directed fragmentation at either amide bond nally to K4 yielding b3 ions, as shown in Figure 5b, except adjacent to the lysine residue. However, after the ϩ ϩ for [Ac-AADKAAAAR H] , in which no b3 or b4 ions crosslinking reaction which produces an amide bond were detected (i.e., peak abundance was below the peak involving the amino group of the lysine side chain, the threshold). For the doubly-charged unmodified peptides, third proton of the crosslinked peptide is likely not as the N-terminal cleavage bias increased over the singly- localized as that of the mobile proton for the unmodi- charged peptides, with the Ac-AADKAAAAR peptide as fied peptides. Ongoing work in our lab examining the lone exception. This particular unmodified peptide infrared multiphoton dissociation of crosslinked pep- exhibited similar cleavage biases for both b- and y-ions. tides has shown enhanced photodissociation efficien- Interestingly, analysis of the b-ions suggests there is a cies for the triply-charged crosslinked peptides com- significant C-terminal cleavage bias for all of the pared to the doubly-charged unmodified peptides, crosslinked peptides. The doubly-charged crosslinked suggesting a lower activation energy for the less basic peptides have both protons localized on the C-terminal crosslinked species. The gas-phase basicity differences arginines, so one would not expect to detect b3 ions, which and probable localization of protons between do not contain the crosslink or a charge. Cleavage C- crosslinked and non-crosslinked peptides may help terminal to the lysine residue yields a b4␣ ion, which explain this enhanced C-terminal cleavage bias to the contains the crosslinker as well as the ␤ peptide with the linked lysine residues. sequestered proton at its C-terminal arginine. A high C-terminal cleavage bias for the formation of b-ions also Dissociation of Other Crosslinked Peptides was observed for all of the triply-charged crosslinked ␣ ␣ peptides. Even with the additional mobile proton, b3␣ ions The peptides -MSH (Ac-SYSMEHFRWGKPV-NH2, were not detected at high abundances in the CAD spectra. peptide) and neurotensin (Pyr-LYENKPRRPYIL, ␤ pep- Having a proline residue N-terminal to the lysine tide) were crosslinked by either DSS or DST to deter- reduces the abundance of N-terminal fragment ions mine whether the same dissociation trends are ob-

(e.g., b3), while a proline C-terminal enhances the served for crosslinked peptides with more random ϩ formation of b4 ions. In the 3 charge state, the aspartic sequences. Since the two crosslinkers react with pri- acid effect does not dominate the peptide fragmentation mary amines, it was assumed that the crosslink was pathways, and any cleavage bias due to the aspartic formed between K11 of ␣-MSH and K6 of neurotensin, acid residues is eliminated. However, the high C- and the product ion spectra of the DST- crosslinked terminal b-ion abundance relative to N-terminal frag- species support this assumption (Figure 6). Dissociation ments for the Asp-containing crosslinked peptides, as of the triply-charged precursor ion was dominated by well as the Ac-AAAKAAAAR peptide, suggests that cleavage C-terminal to the acidic residues, E5 of ␣-MSH 3ϩ 3ϩ the presence of the crosslink affects the dissociation of and E4 of neurotensin, yielding y8␣ (m/z 937.8) and y9␤ the amide bonds adjacent to the modified lysine. (m/z 978.7) as the three protons are localized by the The observed C-terminal cleavage bias could be due to three total arginine residues present in the two peptides the composition and position of the amino acids residues (Figure 6a). To a lesser degree proline-directed cleav- 3ϩ chosen for this study. Cleavage C-terminal to the ages were observed, producing b11␣ (m/z 1080.0) and 3ϩ crosslinked lysine yields the y5␣ fragment ion (XAAAR) b9␤ (m/z 982.7) Without a mobile proton, dissociation of ϩ ϩ ␤ andab4␣ fragment ion (Ac-AAXK crosslinker the C7–C8 bond of the DST crosslink was observed ␤ 2ϩ ␣ ϩ peptide), while N-terminal cleavage produces the b3␣ producing L 7 and the corresponding L 7 product fragment (Ac-AAX) and the corresponding y6␣ fragment ions of m/z 865.6 and 1721.9, respectively. Other ϩ ϩ ␤ (KXAAAR crosslinker peptide). The b3 fragment is crosslinker specific fragment ions were detected includ- unlikely to carry a charge without a basic residue present ing cleavage of the crosslink amide bond yielding the ␤ 2ϩ ␣ ϩ and thus would not be observed. The b4␣,y6␣, and y5␣ unmodified constituent peptides (L 5 and L 5 ), and fragments would all be detected as ions due to the the constituent peptides modified by the DST J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 355

Figure 6. CAD spectra of ␣-MSH crosslinked to neurotensin by DST in the (a)3ϩ charge state and (b)4ϩ charge state. Product ion spectra of ␣-MSH crosslinked to neurotensin by DSS in the (c)3ϩ ϩ ␣ ␣ charge state and (d)4 charge state. -MSH (Ac-SYSMEHFRWGKPV-NH2) is referred to as the peptide and neurotensin (Pyr-LYENKPRRPYIL) as the ␤ peptide; the crosslink is between K11 of ␣-MSH and K6 of neurotensin. An asterisk is used to signify the precursor ion.

ϩ ϩ crosslinker (L␤92 and L␣9 ). Interestingly, crosslinked peptide at similar high abundances as the DST- lysine immonium ions were observed at less than 0.5% crosslinked species without a mobile proton. Such prod- relative abundance for the triply-charged precursor ion, uct ions accounted for less than 1% of the total fragment compared to ϳ5% for the model crosslinked peptide ion abundance. Collisionally activated dissociation of the species without a mobile proton. CAD of the quadru- quadruply-charged DSS-crosslinked peptide yielded ply-charged precursor ion [␣-MSH ϩ DST ϩ neuroten- higher sequence coverage, specifically a series of y-ions of 4ϩ sin ϩ 4H] yielded a full series of crosslinked y␣ ions the ␣-MSH peptide crosslinked to intact neurotensin (Fig-

(y3␣–y12␣), and a full series of uncrosslinked b␣-ions ure 6d). Acidic residue directed cleavages were not ob- (b2␣–b10␣), as shown in Figure 6b. With an additional served for the precursor ion possessing a mobile proton; mobile proton, fragment ions stemming from the cleav- dissociation N-terminal to proline was still dominant age of amide bonds C-terminal to acidic residues were yielding the b11␣ ions. No crosslinker specific product not observed. Proline-directed fragmentation yielded ions, including crosslinked lysine immonium ions, nor 3ϩ the most abundant product ion b11␣. Similar to the CAD fragments from the cleavage of the crosslink amide bonds results of the model crosslinked peptides, crosslinker were observed for this crosslinked peptide ion. specific product ions were not observed at any signifi- cant abundance; instead b- and y-type fragment ions Conclusions dominated the entire CAD spectrum. Most of the prod- uct ions were a result of backbone cleavage along Intermolecularly crosslinked peptides, in general, dis- ␣-MSH, and not neurotensin, with the exception of sociate via similar pathways to that of unmodified amino acid specific fragmentation. peptides. For DSS- and DST-crosslinked peptide ions in The DSS-crosslinked species in the 3ϩ charge state which the protons were sequestered, aspartic acid- produced similar fragment ions as the DST-crosslinked directed fragmentation dominated the product ion analog (Figure 6c). Ions resulting from cleavage C-termi- spectra, more so than fragmentation N-terminally to nal to the two glutamic acid residues dominated the CAD proline residues among the two linked peptides. These 3ϩ 3ϩ spectrum (y8␣ and y9␤ ), and proline-directed products doubly-charged crosslinked peptides yielded a high 3ϩ 3ϩ were also abundant (b11␣ and b9␤ ). Crosslinker specific abundance of multiple cleavage products, including product ions were not observed for the DSS-crosslinked b/b, b/y, y/y, crosslinked lysine immonium ions, and 356 GARDNER AND BRODBELT J Am Soc Mass Spectrom 2008, 19, 344–357 internal ions, which complicated the interpretation of product ion types may lead to incorrect peptide and the CAD spectra, but provided complementary se- protein identifications. Also, these crosslinker specific quence information. In particular, for crosslinked pep- product ion types can provide not only complementary tides with a proline N-terminal to the linked lysine but also confirmatory sequence information which can aid residue (i.e., a PK sequence), a high degree of in identifying the location of the chemical crosslink. crosslinked internal ions were formed. The DST- crosslinked peptide ions without a mobile proton dis- sociated in a unique manner compared to the analogous Acknowledgments DSS species, in which cleavage of the C7–C8 bond of the DST linker molecule was observed regardless of amino The authors gratefully acknowledge funding from the NSF (CHE- acid content. 0718320) and the Welch Foundation (F1155). Dissociation of the crosslinked peptides possessing a mobile proton resulted in almost full sequence coverage (b- and y-type ions) of both peptides, allowing one to References determine the site of crosslinking, as cleavage C-terminal 1. Heck, A. J. R.; van den Heuvel, R. H. H. Investigation of Intact Protein to Asp residues was observed to a much lesser degree. Complexes by Mass Spectrometry. Mass Spectrom. Rev. 2004, 23(5), 368–389. Dissociation N-terminal to the proline residues was still 2. Juan, H.-F.; Liu, H.-L.; Hsu, J.-P. Recent Developments in Structural prevalent for the more highly-charged crosslinked pep- Proteomics: From Protein Identifications and Structure Determinations to Protein-Protein Interactions. Curr. Proteom. 2004, 1(3), 183–197. tides. Products unique to crosslinked peptides, such as 3. Borch, J.; Jorgensen, T. J. D.; Roepstorff, P. Mass Spectrometric Analysis crosslinked lysine immonium ions, multiple cleavage of Protein Interactions. Curr. Opin. Chem. Biol. 2005, 9(5), 509–516. 4. Sinz, A. Chemical Crosslinking and Mass Spectrometry for Mapping fragments, and cleavage of the crosslinkers, including Three-Dimensional Structures of Proteins and Protein Complexes. J. both the crosslinker amide bond and within the DST Mass Spectrom. 2003, 38(12), 1225–1237. 5. Back, J. W.; De Jong, L.; Muijsers, A. O.; De Koster, C. G. Chemical linker, accounted for less than 10% of the fragment ion Cross-Linking and Mass Spectrometry for Protein Structural Modeling. abundance for the triply-charged species, compared to J. Mol. Biol. 2003, 331(2), 303–313. 6. Young, M. M.; Tang, N.; Hempel, J. C.; Oshiro, C. M.; Taylor, E. W.; upwards of 50% of the fragment ions for precursor ion Kuntz, I. D.; Gibson, B. W.; Dollinger, G. High Throughput Protein Fold without a mobile proton. In addition, the crosslinked Identification by Using Experimental Constraints Derived from In- tramolecular Crosslinks and Mass Spectrometry. Proc. Natl. Acad. Sci. peptides exhibited a tendency to preferentially cleave U.S.A. 2000, 97(11), 5802–5806. 7. Huang, B. X.; Kim, H.-Y.; Dass, C. Probing Three-Dimensional Structure C-terminally to the modified lysine residue, particularly in of Bovine Serum Albumin by Chemical Crosslinking and Mass Spec- the formation of b-ions, compared to the unmodified trometry. J. Am. Soc. Mass Spectrom. 2004, 15(8), 1237–1247. 8. Pearson, K. M.; Pannell, L. K.; Fales, H. M. Intramolecular Crosslinking peptides. Unless an aspartic acid residue was positioned Experiments on Cytochrome c and Ribonuclease A Using an Isotope N-terminally to the crosslinked lysine and the precursor Multiplet Method. Rapid Commun. Mass Spectrom. 2002, 16(3), 149–159. 9. Lanman, J.; Lam, T. T.; Barnes, S.; Sakalian, M.; Emmett, M. R.; Marshall, ion did not have a mobile proton, this C-terminal cleavage A. G.; Prevelige, P. E. Identification of Novel Interactions in HIV-1 bias was observed for all of the crosslinked peptides Capsid Protein Assembly by High-resolution Mass Spectrometry. J. Mol. Biol. 2003, 325(4), 759–772. regardless of the position of proline and aspartic acid 10. Back, J. W.; Sanz, M. A.; De Jong, L.; De Koning, L. J.; Nijtmans, L. G. J.; residues. De Koster, C. G.; Grivell, L. A.; Van Der Spek, H.; Muijsers, A. O. A Structure for the Yeast Prohibiting Complex: Structure Prediction and By incorporating these dissociation trends noted for Evidence from Chemical Crosslinking and Mass Spectrometry. Protein crosslinked peptides into computer algorithms, one Sci. 2002, 11(10), 2471–2478. 11. Chu, F.; Shan, S.-O.; Moustakas, D. T.; Alber, F.; Egea, P. F.; Stroud, would expect the confidence in the assignment of product R. M.; Walter, P.; Burlingame, A. L. Unraveling the Interface of Signal ions and subsequent peptide identification to increase. In Recognition Particle and Its Receptor by Using Chemical Crosslinking and Tandem Mass Spectrometry. Proc. Natl. Acad. Sci. U.S.A. 2004, the presence of a mobile proton the intermolecularly 101(47), 16454–16459. crosslinked peptide tends to behave as two individual 12. Kalkhof, S.; Ihling, C.; Mechtler, K.; Sinz, A. Chemical Crosslinking and High-Performance Fourier Transform Ion Cyclotron Resonance Mass peptides, each modified at the lysine residue by the Spectrometry for Protein Interaction Analysis: Application to a Calmod- ulin/Target Peptide Complex. Anal. Chem. 2005, 77(2), 495–503. crosslinker and the other peptide. However, as evident in 13. Ahrends, R.; Kosinski, J.; Kirsch, D.; Manelyte, L.; Giron-Monzon, L.; dissociation of the crosslinked model peptides without a Hummerich, L.; Schulz, O.; Spengler, B.; Friedhoff, P. Identifying an Interaction Site Between MutH and the C-Terminal Domain of MutL by mobile proton, several unique types of fragment ions were Crosslinking, Affinity Purification, Chemical Coding, and Mass Spec- identified. Many commercial crosslinkers are primary trometry. Nucleic Acids Res. 2006, 34(10), 3169–3180. 14. Huang, B. X.; Dass, C.; Kim, H.-Y. Probing Conformational Changes of amine specific, and after crosslinking, in which an amide Human Serum Albumin Due to Unsaturated Fatty Acid Binding by bond is formed, the overall gas-phase basicity of the Chemical Crosslinking and Mass Spectrometry. Biochem. J. 2005, 387(3), 695–702. peptide or protein is reduced, thus decreasing the ioniza- 15. Huang, B. X.; Kim, H.-Y. Interdomain Conformational Changes in Akt tion efficiency of more highly-charged crosslinked pep- Activation Revealed by Chemical Crosslinking and Tandem Mass Spectrometry. Mol. Cell. Proteom. 2006, 5(6), 1045–1053. tides. Particularly for tryptic peptides in which basic 16. Urlaub, H.; Hartmuth, K.; Kostka, S.; Grelle, G.; Luhrmann, R. A residues are not conserved, crosslinked peptides without a General Approach for Identification of RNA-Protein Crosslinking Sites Within Native Human Spliceosomal Small Nuclear Ribonucleoproteins mobile proton dissociate via several pathways yielding (snRNPs). Analysis of RNA-Protein Contacts in Native U1 and U4/ multiple cleavage product ions (e.g., b␣y␤), crosslinked U6.U5 snRNPs. J. Biol. Chem. 2000, 275(52), 41458–41468. 17. Chen, T.; Jaffe, J. D.; Church, G. M. Algorithms for Identifying Protein lysine immonium ions, and fragments from the dissocia- Crosslinks Via Tandem Mass Spectrometry. J. Comput. Biol. 2001, 8(6), tion of the crosslink amide bond or of the crosslinker (e.g., 571–583. 18. de Koning, L. J.; Kasper, P. T.; Back, J. W.; Nessen, M. A.; Vanrobaeys, L␣7 and L␤7 fragments of DST crosslinked peptides), as F.; Van Beeumen, J.; Gherardi, E.; de Koster, C. G.; de Jong, L. Computer-Assisted Mass Spectrometric Analysis of Naturally Occur- well as the typical b- and y-type product ions. Computer ring and Artificially Introduced Crosslinks in Proteins and Protein algorithms which do not currently account for these Complexes. FEBS J. 2006, 273(2), 281–291. J Am Soc Mass Spectrom 2008, 19, 344–357 EFFECTS OF Asp/Pro ON CROSSLINKED PEPTIDES 357

19. Gao, Q.; Xue, S.; Doneanu, C. E.; Shaffer, S. A.; Goodlett, D. R.; Nelson, Containing Proteins from Electrospray Ionization. Anal. Chem. 1993, S. D. Pro-CrossLink Software Tool for Protein Cross-Linking and Mass 65(4), 425–438. Spectrometry. Anal. Chem. 2006, 78(7), 2145–2149. 33. Tsaprailis, G.; Somogyi, A.; Nikolaev, E. N.; Wysocki, V. H. Refining the 20. Schilling, B.; Row, R. H.; Gibson, B. W.; Guo, X.; Young, M. M. Model for Selective Cleavage at Acidic Residues in Arginine-Containing MS2Assign, Automated Assignment and Nomenclature of Tandem Protonated Peptides. Int. J. Mass Spectrom. 2000, 195/196, 467–479. Mass Spectra of Chemically Crosslinked Peptides. J. Am. Soc. Mass 34. Papayannopoulos, I. A. The Interpretation of Collision-Induced Disso- Spectrom. 2003, 14(8), 834–850. ciation Tandem Mass Spectra of Peptides. Mass Spectrom. Rev. 1995, 21. Seebacher, J.; Mallick, P.; Zhang, N.; Eddes, J. S.; Aebersold, R.; Gelb, 14(1), 49–73. M. H. Protein Crosslinking Analysis Using Mass Spectrometry, Isotope- 35. Reid, G. E.; Wu, J.; Chrisman, P. A.; Wells, J. M.; McLuckey, S. A. Charge Coded Cross-Linkers, and Integrated Computational Data Processing. J. State-Dependent Sequence Analysis of Protonated Ubiquitin Ions Via Proteome Res. 2006, 5(9), 2270–2282. Ion Trap Tandem Mass Spectrometry. Anal. Chem. 2001, 73(14), 3274– 22. Tang, Y.; Chen, Y.; Lichti, C. F.; Hall, R. A.; Raney, K. D.; Jennings, S. F. 3281. CLPM: A Crosslinked Peptide Mapping Algorithm for Mass Spectro- 36. Reid, G. E.; Stephenson, J. L., Jr.; McLuckey, S. A. Tandem Mass metric Analysis. BMC Bioinformatics 2005, 6(Suppl. 2), S9. Spectrometry of Ribonuclease A and B: N-Linked Glycosylation Site 23. Gaucher, S. P.; Hadi, M. Z.; Young, M. M. Influence of Crosslinker Analysis of Whole Protein Ions. Anal. Chem. 2002, 74(3), 577–583. Identity and Position on Gas-Phase Dissociation of Lys-Lys Crosslinked 37. Sullivan, A. G.; Brancia, F. L.; Tyldesley, R.; Bateman, R.; Sidhu, K.; Peptides. J. Am. Soc. Mass Spectrom. 2006, 17(3), 395–405. Hubbard, S. J.; Oliver, S. G.; Gaskell, S. J. The Exploitation of Selective 24. Raftery, M. J.; Geczy, C. L. Electrospray Low Energy CID and MALDI Cleavage of Singly-Protonated Peptide Ions Adjacent to Aspartic Acid PSD Fragmentations of Protonated Sulfinamide Crosslinked Peptides. Residues Using a Quadrupole Orthogonal Time-of-Flight Mass Spec- J. Am. Soc. Mass Spectrom. 2002, 13(6), 709–718. trometer Equipped with a Matrix-Assisted Laser Desorption/Ionization 25. Hunt, D. F.; Yates, J. R., III; Shabanowitz, J.; Winston, S.; Hauer, C. R. Source. Int. J. Mass Spectrom. 2001, 210/211(1/3), 665–676. 38. Wysocki, V. H.; Tsaprailis, G.; Smith, L. L.; Breci, L. A. Mobile and Protein Sequencing by Tandem Mass Spectrometry. Proc. Natl. Acad. Sci. Localized Protons: A Framework for Understanding Peptide Dissocia- U.S.A. 1986, 83(17), 6233–6237. tion. J. Mass Spectrom. 2000, 35(12), 1399–1406. 26. Biemann, K.; Martin, S. A. Mass Spectrometric Determination of the 39. Burlet, O.; Orkiszewski, R. S.; Ballard, K. D.; Gaskell, S. J. Charge Amino Acid Sequence of Peptides and Proteins. Mass Spectrom. Rev. Promotion of Low-Energy Fragmentations of Peptide Ions. Rapid Com- 1987, 6(1), 1–76. mun. Mass Spectrom. 1992, 6(11), 658–662. 27. Huang, Y.; Triscari, J. M.; Tseng, G. C.; Pasa-Tolic, L.; Lipton, M. S.; 40. Summerfield, S. G.; Gaskell, S. J. Fragmentation Efficiencies of Peptide Smith, R. D.; Wysocki, V. H. Statistical Characterization of the Charge Ions Following Low Energy Collisional Activation. Int. J. Mass Spectrom. State and Residue Dependence of Low-Energy CID Peptide Dissocia- Ion Processes 1997, 165/166, 509–521. tion Patterns. Anal. Chem. 2005, 77(18), 5800–5813. 41. Harrison, A. G.; Tn, Y.-P. Ion Chemistry of Protonated Aspartic Acid 28. Tabb, D. L.; Smith, L. L.; Breci, L. A.; Wysocki, V. H.; Lin, D.; Yates, J. R., Derivatives. J. Mass Spectrom. 1998, 33(6), 532–542. III. Statistical Characterization of Ion Trap Tandem Mass Spectra from 42. Yu, W.; Vath, J. E.; Huberty, M. C.; Martin, S. A. Identification of the Doubly-Charged Tryptic Peptides. Anal. Chem. 2003, 75(5), 1155–1163. Facile Gas-Phase Cleavage of the Asp-Pro and Asp-Xxx Peptide Bonds 29. Kapp, E. A.; Schutz, F.; Reid, G. E.; Eddes, J. S.; Moritz, R. L.; O’Hair, in Matrix-Assisted Laser Desorption Time-of-Flight Mass Spectrometry. R. A.; Speed, T. P.; Simpson, R. J. Mining a Tandem Mass Spectrometry Anal. Chem. 1993, 65(21), 3015–3023. Database to Determine the Trends and Global Factors Influencing 43. Gu, C.; Tsaprailis, G.; Breci, L.; Wysocki, V. H. Selective Gas-Phase Peptide Fragmentation. Anal. Chem. 2003, 75(22), 6251–6264. Cleavage at the Peptide Bond C-Terminal to Aspartic Acid in Fixed- 30. Dongre, A. R.; Jones, J. L.; Somogyi, A.; Wysocki, V. H. Influence of Charge Derivatives of Asp-Containing Peptides. Anal. Chem. 2000, Peptide Composition, Gas-Phase Basicity, and Chemical Modification 72(23), 5804–5813. on Fragmentation Efficiency: Evidence for the Mobile Proton Model. 44. Loo, J. A.; Edmonds, C. G.; Smith, R. D. Primary Sequence Information J. Am. Chem. Soc. 1996, 118(35), 8365–8374. from Intact Proteins by Electrospray Ionization Tandem Mass Spec- 31. Tsaprailis, G.; Nair, H.; Somogyi, A.; Wysocki, V. H.; Zhong, W.; Futrell, trometry. Science 1990, 248(4952), 201–204. J. H.; Summerfield, S. G.; Gaskell, S. J. Influence of Secondary Structure 45. Schwartz, B. L.; Bursey, M. M. Some Proline Substituent Effects in the on the Fragmentation of Protonated Peptides. J. Am. Chem. Soc. 1999, Tandem Mass Spectrum of Protonated Penta-Alanine. Biol. Mass Spec- 121(22), 5142–5154. trom. 1992, 21(2), 92–96. 32. Loo, J. A.; Edmonds, C. G.; Smith, R. D. Tandem mass spectrometry of 46. Vaisar, T.; Urban, J. Probing the Proline Effects in CID of Protonated very large molecules. II. Dissociation of Multiply-Charged Proline- Peptides. J. Mass Spectrom. 1996, 31(10), 1185–1187.