Statistical and Mechanistic Approaches to ... - ACS Publications

Statistical and Mechanistic Approaches to Understanding the Gas-Phase Fragmentation Behavior of Methionine Sulfoxide Containing Peptides Gavin E. Reid,* Kade D. Roberts, Eugene A. Kapp, and Richard J. Simpson Joint ProteomicS Laboratory, Ludwig Institute for Cancer Research and The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia Received January 21, 2004

Recently, we carried out a statistical analysis of a ‘tryptic’ peptide tandem mass spectrometry database in order to identify sequence-dependent patterns for the gas-phase fragmentation behavior of protonated peptide ions, and to improve the models for peptide fragmentation currently incorporated into peptide sequencing and database search algorithms [Kapp, E. A., Schu¨ tz, F., Reid, G. E., Eddes, J. S., Moritz, R. L., O’Hair, R. A. J., Speed, T. P. and Simpson, R. J. Anal. Chem. 2003, 75, 6251-6264.]. Here, we have reexamined this database in order to determine the effect of a common post-translational or process induced modification, methionine oxidation, on the appearance and relative abundances of the product ions formed by low energy collision induced dissociation of peptide ions containing this modification. The results from this study indicate that the structurally diagnostic neutral loss of methane sulfenic acid (CH3SOH, 64Da) from the side chain of methionine sulfoxide residues is the dominant fragmentation process for methionine sulfoxide containing peptide ions under conditions of low proton mobility, i.e., when ionizing proton(s) are sequestered at strongly basic amino acids such as arginine, lysine or histidine. The product ion abundances resulting from this neutral loss were found to be approximately 2-fold greater than those resulting from the cleavage C-terminal to aspartic acid, which has previously been shown to be enhanced under the same conditions. In close agreement with these statistical trends, experimental and theoretical studies, employing synthetic “tryptic” peptides and model methionine sulfoxide containing peptide ions, have determined that the mechanism for enhanced methionine sulfoxide side chain cleavage proceeds primarily via a ‘charge remote’ process. However, the mechanism for dissociation of the side chain for these ions was observed to change as a function of proton mobility. Finally, the transition state barrier for the charge remote side chain cleavage mechanism is predicted to be energetically more favorable than that for charge remote cleavage C-terminal to aspartic acid. Keywords: mass spectrometry • methionine oxidation • peptide fragmentation

Introduction The presence of a post-translationally modified (PTM) amino acid residue within a peptide sequence often has a dramatic effect on the fragmentation behavior of its protonated ions under low energy collision induced dissociation (CID) tandem mass spectrometry (MS/MS) conditions.1 In many instances, the formation of ‘nonsequence’ product ions, such as those corresponding to cleavage at the side chain of an amino acid residue containing a PTM, may dominate the product ion spectrum, while ‘sequence’ product ions, formed by amide bond cleavages along the peptide backbone, are suppressed.2-10 Although nonsequence side chain cleavage product ions are diagnostic for the presence of the PTM residue within the peptide sequence, their formation in high relative abundance often results in the product ion spectrum being unsuitable for sequence analysis and subsequent identification by manual interpretation,11,12 or by database interrogation.13-16 As current proteomics strategies are becoming increasingly reliant on the * To whom correspondence should be addressed. Joint ProteomicS Laboratory, Ludwig Institute for Cancer Research, P.O. Box 2008 Royal Melbourne Hospital, Parkville, Vic., Australia 3052. Ph: +61-3-9341-3155. Fax: +61-3-9341-3192. E-mail: [email protected] 10.1021/pr0499646 CCC: $27.50

 2004 American Chemical Society

use of automated database search algorithms for identifying proteins from peptide MS/MS data, it is expected that determination of the factors that influence the formation of ‘nonsequence’ versus ‘sequence’ ions from the MS/MS of posttranslationally modified peptide ions, and incorporation of the fragmentation “rules” arising from these studies into the database search algorithms, would aid in the development of more effective tools for high throughput protein identification. Recently, we have carried out a statistical analysis of a database of 5,500 unique ‘tryptic’ peptide tandem mass spectra acquired in an ion trap mass spectrometer, to determine the global factors influencing the sequence-dependent gas-phase fragmentation behavior of protonated peptide ions under low energy CID conditions.17 In this study, cleavage intensity ratio’s (CIR), calculated by dividing the summed product ion abundances of the complementary N-terminal b-type and/or C-terminal y-type product ions observed at each cleavage site by the average abundance of all cleavage sites within the peptide, were determined in order to quantitatively measure the extent of ‘sequence’ ion fragmentation occurring at each site along the peptide backbone. CIR values for fragmentations occurring C-terminal to Aspartic acid within peptide ions containing AspJournal of Proteome Research 2004, 3, 751-759

751

Published on Web 04/29/2004

research articles Xaa bonds, previously demonstrated to occur via a ‘charge remote’ process (eq 1)18-25 when ionizing protons are ‘sequestered’ at the strongly basic side chains of arginine, lysine or histidine residues contained within the peptide sequence, i.e., under conditions of low proton mobility,23,26 were then used to derive a ‘relative proton mobility’ scale classification scheme for all peptide spectra contained within the database. Note that the term ‘charge remote’27 is used here to denote fragmentation reactions that are not catalyzed by an ionizing proton; i.e., the charge acts as a spectator.

This classification scheme has provided the basis upon which to quantitatively determine the influence of each amino acid residue N- or C-terminal to a given amide bond cleavage site along the peptide backbone, as well as the positional effect of the cleavage site, on the appearance and relative abundances of the resultant product ion spectrum. The results from this analysis have been employed to re-evaluate currently acceptable thresholds for MS/MS search algorithm cutoff filters,17 and are being used to develop improved scoring functions for MS/ MS search algorithms, as well as to develop predictive models of peptide fragmentation for de novo sequence analysis.28 It is important to note, however, that this previous study only took into consideration the b- and y-type ‘sequence’ ions resulting from fragmentation of peptide amide bonds within the amino acid sequence. Therefore, the contribution of ‘nonsequence’ ion fragmentation pathways, such as those resulting in the formation of a-type ions by the loss of CO from b-type ions,29-31 the loss of small molecules such as NH3 or H2O from amino acid side chains or the peptide backbone,2,32-38 the loss of diagnostic side chain fragments from post-translationally modified amino acid residues,2-10,39 sequential fragmentations,40 or intramolecular rearrangement product ions,41-43 to the resultant product ion spectrum were not evaluated. Aside from a recent report from the laboratories of Wysocki and Yates,44 who have examined factors influencing the appearance of the ‘nonsequence’ ion neutral losses of NH3 or H2O from the various amide backbone cleavage product ions within a limited database of doubly protonated tryptic peptides, there have been no systematic studies carried out to quantitatively determine the global effects of nonsequence ion fragmentation pathways, and particularly those involving post-translationally modified amino acid residues, on the gas-phase fragmentation behavior of protonated peptide ions. Methionine oxidation is one of the more common modifications encountered during the sequence analysis of proteins by 752

Journal of Proteome Research • Vol. 3, No. 4, 2004

Reid et al.

tandem mass spectrometry (MS/MS).3-7 While the oxidation of methionine is often considered to be an artifact of the gel electrophoresis sample preparation methods typically employed for protein purification, a number of in vivo studies have revealed a key biological role for methionine oxidation in molecular mechanisms for the regulation of cellular function and in defense against reactive oxygen species-mediated protein damage, as well as being apparent in pathologies associated with age-related or degenerative diseases such as Alzheimer’s.45-48 The identification of peptide ions containing methionine sulfoxide, as well as the specific site of the oxidized methionine residue within the peptide sequence, may be facilitated by the diagnostic loss of methane sulfenic acid (CH3SOH, 64 Da) from either the precursor or product ions, respectively, under both low and high energy CID MS/MS conditions.3-7,49-52 Recognition of this diagnostic neutral loss is of particular importance given that methionine sulfoxide and phenylalanine residues have the same nominal mass. Previously, it has been observed that the loss of CH3SOH from methionine sulfoxide residues occurs predominantly from singly protonated tryptic peptide ions:3-7 i.e., under conditions of low proton mobility similar to those previously described for enhanced cleavage occurring at Asp-Xaa bonds, as discussed above. When this ‘nonsequence’ loss is observed as the dominant fragmentation process within a given peptide ion, the formation of ‘sequence’ ions are commonly suppressed, such that the ability to identify the peptide is hindered. On the basis of these observations, we have re-examined our peptide MS/MS database in order to systematically determine the influence of factors such as charge state and amino acid composition on the low energy collision induced dissociation of peptide ions containing methionine sulfoxide, and the effect of this modification on the appearance and relative abundances of the resultant product ion spectra. Furthermore, we have examined the fragmentation behavior of several synthetic model “tryptic” peptides containing methionine sulfoxide, by using tandem mass spectrometry and hydrogen/deuterium exchange, and employed molecular orbital calculations on model methionine sulfoxide containing peptide ions, to determine the mechanisms for enhanced methionine sulfoxide side chain cleavage, and to provide a sound chemical rationale for the observed statistical trends.

Experimental Section Materials. The synthetic “tryptic” peptides GAILMGAILR and GAILMGAILK were obtained from Auspep (Melbourne, Vic). Methanol (ChromAR grade) was purchased from Mallinckrodt (Melbourne, Australia). Glacial acetic acid was purchased from BDH laboratories (Poole, England). Monodeuterated methanol (CH3OD) and acetic acid (CH3CO2D) were from Aldrich (Milwaukee, WI). Deuterium oxide (D2O) was obtained from Cambridge Isotope Laboratories (Andover, MA). Hydrogen peroxide was from Merck (Darmstadt, Germany). All reagents were used without further purification. Peptide MS/MS Database. The construction and content of the peptide CID MS/MS spectra database was described recently.17 Briefly, peptide spectra were acquired by capillary RP-HPLC CID MS/MS, following in-gel tryptic digestion of proteins resolved by one or two-dimensional gel electrophoresis,4,6,8 using a quadrupole ion trap mass spectrometer (model LCQ; Finnigan, San Jose, CA) equipped with electrospray ionization (ESI). The amino acid sequences of uninterpreted CID MS/MS spectra were determined by correlation with

research articles

Gas-Phase Fragmentation Behavior of Met(O)

predicted spectra of peptide sequences present in a nonredundant protein sequence database (950 000 proteins)53 using the SEQUEST algorithm,13-15 then manually validated using an interactive in-house program, “CHOMPER”53 prior to database submission. The database contains 5500 unique spectra (i.e., where duplicate spectra of the same charge state but lower SEQUEST cross-correlation scores have been removed). Statistical Analysis Using Cleavage Intensity Ratio (CIR) Calculations. Spectra were classified according to the charge state and amino acid composition of the precursor ion, as ‘mobile’ when the total number of protons > number of basic residues (i.e., combined number of Arg, Lys and His residues), ‘partially mobile’ when the combined number of basic residues g total number of protons > number of Arg residues, or ‘nonmobile’ when the total number of protons e number of Arg residues,17 thus taking into account the differences in proton affinities and ability to ‘sequester’ protons of these basic amino acids. Then, to quantify the extent of fragmentation occurring at the side chain of methionine sulfoxide residues, or C-terminal to aspartic acid, cleavage intensity ratio’s (CIR) were determined by dividing either the abundance of the methionine sulfoxide side chain cleavage product ion, or the summed b- and y-ion abundances for Asp-Xaa cleavage, respectively, by the average abundance of all cleavage sites within the peptide, as shown in eq 2 z

∑b

z+ s

+ y z+ s or Met(O) loss

z)1

CIRs ) 1

N

(2)

z

∑ ∑(b N

z+ i

+ y z+ i + Met(O) loss)

i)1 z)1

where N ) number of cleavage sites, s ) cleavage site of interest (1 e s e N), z ) charge state of the precursor ion, bz+i and yz+i are the normalized abundances of the b- and y-type ions with z charge at the ith cleavage site, respectively, and Met(O) loss is the normalized abundance of the methionine sulfoxide side chain neutral loss. Average CIR values for both these cleavages were then calculated for each proton mobility classification, using the individual CIR values of the peptides contained within each classification. Average CIR values greater than 1.0 indicate enhanced cleavages. All CIR values were calculated based on a “no cutoff” data set, whereby CIR values were determined for all amide bond cleavage sites, regardless of whether one of the potential product ions for that cleavage site fell below the low mass cut off (LMCO) of the instrument. Methionine Sulfoxide Formation. Methionine-containing synthetic peptides (10 µg) were dissolved in 50 µL of 30% hydrogen peroxide containing 5% acetic acid and allowed to react at room temperature for 15 min. Samples were then dried by lyophilization prior to being dissolved for mass spectrometric analysis. Mass Spectrometry. Singly protonated ([M+H]+) and doubly protonated ([M+2H]2+) precursor ions were formed via electrospray ionization (ESI) on a Finnigan model LCQ-deca (San Jose, CA) quadrupole ion trap mass spectrometer. Samples, (0.01 mg/mL) dissolved in 50% CH3OH/50% H2O containing 0.1M acetic acid were introduced to the mass spectrometer at 2 µL/min. The spray voltage was set at 4.5 kV. Nitrogen sheath gas was supplied at 25 psi. The heated capillary temperature was 200 °C. Deuterium labeled peptides were prepared by dissolving the sample (0.01 mg/mL) in 50%CH3OD/50%D2O

containing 0.1M CH3CO2D. [M+D]+ and [M+2D]2+ ions were then introduced to the mass spectrometer under the same conditions as described above, with the addition of an auxiliary gas (15, arbitrary units) to reduce deuterium back exchange during electrospray sample introduction. CID MS/MS experiments were performed on monoisotopically mass selected ions using standard isolation and excitation procedures. Computational Methods. Transition state structures were initially examined at the PM3 semiempirical level of theory, followed by further optimization at the B3LYP level of theory using the 6-31+G** basis set54 All calculations have been performed using the GAUSSIAN 98 molecular modeling package.55 All optimized structures were subjected to harmonic vibrational frequency analysis and visualized using the computer package GaussView 3.0 to determine the nature of the stationary points.56 Zero-point energies were obtained from harmonic frequency calculations without scaling. Intrinsic reaction coordinate (IRC) runs were performed on each transition state, followed by geometry optimizations to locate the appropriate reactant and product ion structures associated with each transition state. When comparing the relative transition state barriers associated with two different fragmentation pathways (i.e., methionine sulfoxide side chain cleavage versus cleavage C-terminal to aspartic acid), further conformational searches were carried out to locate the lowest energy precursor ion structure in each case. Complete structural details and lists of vibrational frequencies for each B3LYP/6-31+G** optimized structure are available from the authors upon request.

Results and Discussion Statistical Analysis of Product Ion Abundances Corresponding to Methionine Sulfoxide Side Chain Cleavage in Protonated Peptide Ions. To obtain a quantitative assessment of the product ion abundances corresponding to the loss of CH3SOH from the side chain of methionine sulfoxide containing peptide ions, as well as to determine the influence of charge state and amino acid composition on this ‘nonsequence’ ion side chain dissociation pathway under low energy ion trap CID MS/MS conditions, the CIR values for all peptides in the MS/MS database containing methionine sulfoxide were calculated, using the extended CIR equation described in the Experimental section (eq 2). CIR values for product ions corresponding to Asp-Xaa bond cleavages, which, along with product ions corresponding to fragmentations occurring between Pro-Xaa bonds, have been demonstrated to be enhanced upon dissociation of protonated peptide ions under certain conditions,17 were also determined for all peptides in the peptide MS/MS database containing aspartic acid, to allow a comparison of the methionine sulfoxide side chain cleavage CIR values with those data obtained previously. Note that the values for Asp-Xaa cleavage obtained here will be somewhat different from our previous report, due to the incorporation of methionine sulfoxide side chain cleavage product ion abundances in the CIR equation. The average CIR values for each of these cleavages were then calculated and are shown in Table 1, classified according to the relative proton mobility of the peptide ion as ‘nonmobile’, ‘partially mobile’, or ‘mobile’ depending on the charge state and number of basic amino acid residues (arginine, lysine and histidine) contained in the peptide sequence. An examination of Table 1 reveals that the CIR values for methionine sulfoxide side chain cleavage are enhanced under nonmobile and partially mobile conditions, similar to those for Asp-Xaa cleavage. Furthermore, the CIR Journal of Proteome Research • Vol. 3, No. 4, 2004 753

research articles

Reid et al.

Table 1. Average Cleavage Intensity Ratio (CIR) Values for Methionine Sulfoxide (Met(O)) Side Chain Cleavage versus Asp-Xaa Amide Bond Cleavage ‘mobile’ peptidesa charge state of precursor ion

Met(O) containing peptides

singly charged peptides (1+) doubly charged peptides (2+) triply charged peptides (3+)

‘partially mobile’ peptidesb

Asp containing peptides

Met(O) containing peptides


‘nonmobile’ peptidesc Met(O) containing peptides


N/A

N/A

7.75(70)d

2.03(296)

7.56(15)

4.63(171)

0.52(251)

0.89(701)

3.76(400)

1.74(1382)

9.53(28)

4.44(163)

0.33(84)

1.0(230)

1.28(129)

1.59(493)

6.72(4)

3.00(16)

a ‘Mobile’ peptides: total number of protons > number of basic residues (i.e., combined number of Arg, Lys, and His residues). b ‘Partially mobile’ peptides: combined number of basic residues g total number of protons > number of Arg residues. c ‘Nonmobile’ peptides: total number of protons e number of Arg residues. d Number of peptides in each category are shown in parentheses.

Table 2. Average Cleavage Intensity Ratio (CIR) Values for Methionine Sulfoxide Side Chain Cleavage versus Xaa-Asp Amide Bond Cleavage for Peptides Containing Both Methionine Sulfoxide and Aspartic Acid ‘mobile’ peptidesa

singly charged peptides (1+)

N/A

doubly charged peptides (2+)

Met 0.50 Asp 0.90 (163) Met 0.45 Asp 0.98 (52)

triply charged peptides (3+)

‘partially mobile’ peptidesb

‘nonmobile’ peptidesc

Met 8.67 Asp 0.79 (30)d Met 3.92 Asp 1.65 (285) Met 1.31 Asp 1.79 (110)

Met 8.76 Asp 0.78 (11) Met 9.35 Asp 2.67 (20) Met 6.71 Asp 3.63 (4)

a ‘Mobile’ peptides: total number of protons > number of basic residues (i.e., combined number of Arg, Lys, and His residues). b ‘Partially mobile’ peptides: combined number of basic residues g total number of protons > number of Arg residues. c ‘Nonmobile’ peptides: total number of protons e number of Arg residues. d Number of peptides in each category are shown in parentheses.

values for methionine sulfoxide cleavage in the singly, doubly, and triply charged nonmobile proton datasets, as well as the singly and doubly charged partially mobile proton datasets, were found to be approximately two to three times that for Asp-Xaa cleavage, suggesting that methionine sulfoxide side chain cleavage has a more dominant effect on the appearance of the resultant product ion spectra under these conditions. This effect is even more apparent from the data shown in Table 2, where the average CIR values obtained for methionine sulfoxide side chain cleavage in peptides containing both methionine sulfoxide and aspartic acid residues were found to be up to eleven times greater than that for Asp-Xaa cleavage for peptides within the same proton mobility classification. There are a total of 1706 methionine containing peptides (2144 methionine residues) contained within the database, of which 981 (1206 methionine residues) (approximately 57%) were identified in their oxidized form. The relative ease by which methionine is oxidized, as well as the magnitude of the methionine sulfoxide side chain fragmentation compared to Asp-Xaa cleavage, highlights the need to more fully understand the mechanisms and factors affecting this fragmentation, in order for it to be more comprehensively accounted for in database search strategies employed for protein identification. Determining Mechanisms for the Loss of Methane Sulfenic Acid from the Side Chain of Protonated Methionine Sulfoxide Containing Peptide Ions. To determine the mechanisms asso754


ciated with the enhanced loss of CH3SOH from the side chain of methionine sulfoxide residues under low proton mobility conditions, we have examined the fragmentation reactions of two synthetic “tryptic” peptides GAILMGAILR and GAILMGAILK. According to the relative proton mobility scale discussed above, the singly protonated ions of GAILMGAILR and GAILMGAILK are classified as nonmobile and partially mobile, respectively, and their doubly protonated ions classified as mobile. Therefore, the magnitude of the methionine sulfoxide side chain loss from these peptide ions, and the mechanisms responsible for their fragmentation, should be directly related to the statistical trends observed from the peptide MS/MS database. As previously noted,52 the loss of CH3SOH from the side chain of methionine sulfoxide residues can occur in two ways. A charge remote mechanism (eq 3), occurring via a cis-1,2 elimination transition state pathway, would involve the loss of a methylene hydrogen atom from the methionine side chain. In contrast, charge directed neighboring group participation mechanisms, involving nucleophilic attack by either the Nterminal amide carbonyl oxygen (eq 4) or the C-terminal amide carbonyl oxygen (eq 5) on the methionine sulfoxide side chain, would require one of the ionizing protons to be localized at the site of cleavage. Distinguishing these charge-remote versus charge-directed mechanisms from each other may be achieved by examination of the methionine sulfoxide side chain losses occurring under hydrogen/deuterium exchange conditions, where all acidic hydrogens are exchanged for deuterium. The charge remote cis-1,2 elimination reaction would result in the loss of CH3SOH (eq 3), while neighboring group participation fragmentation reactions would result in the loss of CH3SOD (eq 4 and eq 5).


The product ion spectra of oxidized GAILMGAILR in its singly and doubly protonated forms, following hydrogen/deuterium exchange and CID MS/MS, are shown in Figure 1, parts A and B, respectively. The m/z regions around the expected methionine sulfoxide side chain cleavage product ion for the singly and doubly protonated ions, obtained in a high resolution ‘zoom scan’ mode experiment, are shown in Figure 1, parts C and D, respectively. The singly protonated precursor ion of oxidized GAILMGAILR was observed to fragment almost exclusively via the loss of CH3SOH (64 Da) from the side chain of methionine sulfoxide, indicative of the charge remote cis-1,2 elimination mechanism described in eq 3. In contrast, the loss of methane sulfenic acid from the doubly protonated ion was observed as only a minor product, and occurred primarily via a charge directed fragmentation pathway, as evidenced by the loss of CH3SOD (65 Da, observed as the loss of 32.5 Da from the doubly protonated ion) (Figure 1, parts B and D). This general fragmentation behavior was also observed upon dissociation of the singly and doubly charged precursor ions of oxidized GAILMGAILK (Figure 2, parts A and B, respectively). Thus, the mechanism for dissociation of the side chain for these ions was observed to change as a function of proton mobility i.e, charge remote fragmentation for ‘nonmobile’ and ‘partially mobile’ peptide ions, and charge directed fragmentation for ‘mobile’ peptide ions. The absence of product ions other than that corresponding to CH3SOH loss from dissociation of the singly protonated oxidized GAILMGAILR (Figure 1A) is consistent with the ionizing proton (or deuteron) being strongly sequestered at the C-terminal arginine side chain, limiting proton mobility and thereby inhibiting dissociation along the peptide backbone via charge directed fragmentation processes. In contrast, a number of additional product ions were observed for dissociation of the singly protonated oxidized GAILMGAILK ion, albeit at low abundance, indicative of the greater degree of proton mobility for this partially mobile peptide ion compared to that for the nonmobile arginine containing peptide of the same charge state. Overall, the fragmentation behavior and relative abundances of the product ions corresponding to methionine sulfoxide side chain cleavage in the synthetic tryptic peptide ions were consistent with those observed for peptides with the same relative proton mobility classifications contained within the MS/MS database, as shown in Table 1, suggesting that the general mechanism for enhanced methionine sulfoxide side chain cleavage under conditions of low proton mobility occurs predominantly via a charge remote process. When methionine sulfoxide side chain cleavage was observed as the dominant fragmentation pathway via the charge remote mechanism, i.e., from the singly protonated GAILMGAILR and GAILMGAILK peptide ions, the collision energy required for dissociation was observed to be lower than that for dissociation

research articles of the non-oxidized methionine containing peptide ions (Figure 3, parts A and B). (Dissociation of the non-oxidized singly protonated GAILMGAILR and GAILMGAILK peptide ions was observed to predominantly result in the formation of a series of b- and a-type product ions (b9, a9, b8, a8, and b7 and a7), each accompanied by the loss of small molecules of H2O or NH3, respectively (data not shown).) As the limiting step to dissociation along the peptide backbone via charge directed fragmentation processes under nonmobile conditions is expected to be a relatively high energy intramolecular proton transfer from the arginine or lysine side chains to the site of cleavage, this result is consistent with the introduction of a lower energy process upon methionine oxidation under conditions of low proton mobility, i.e., the charge remote side chain fragmentation, compared to the non-oxidized peptides. In contrast, for methionine sulfoxide containing peptides that exhibited only minor side chain cleavage via charge directed fragmentation pathways i.e., the doubly protonated GAILMGAILR and GAILMGAILK peptide ions, an increase in the collision energy required for dissociation was observed compared to the non-oxidized methionine containing peptide ions (Figure 3, parts A and B, respectively). (The product ion abundances and sites of amide bond cleavage for dissociation of these non-oxidized peptide ions were essentially the same as those observed in the oxidized species (data not shown).) This is consistent with an increase in the proton affinity for the methionine sulfoxide side chain compared to that of methionine, thereby resulting in the ionizing protons becoming “less mobile”. Although the absolute values are likely to be higher, the proton affinity of the methionine sulfoxide side chain compared to methionine may be estimated from the proton affinities of the model systems, dimethyl sulfoxide ((CH3)2SO) 211.4 kcal mol-1) and dimethyl sulfide ((CH3SCH3), 198.6 kcal mol-1), respectively. Additionally, we are currently determining an experimental proton affinity value for methionine sulfoxide (preliminary results indicate a value of approximately 232 kcal mol-1,57 compared to the literature proton affinity value for methionine of 223.6 kcal mol-1), as well as a number of other common PTM’s, using the extended kinetic method,58-60 to more fully define the relative proton mobility scale classification approach to statistical analysis. Examining Mechanisms for Methane Sulfenic Acid Loss Compared to Asp-Xaa Cleavage by Gas-Phase Molecular Orbital Calculations. The data shown in Figures 1-3 indicates that the enhanced loss of methane sulfenic acid from methionine sulfoxide containing peptides occurs via an energetically favored charge remote mechanism under nonmobile and partially mobile conditions, while the same loss under mobile conditions is only observed as a minor process and occurs primarily via a charge directed mechanism. To obtain further insights into the relative energies associated with these processes, the transition state barriers for the loss of CH3SOH from methionine sulfoxide containing peptides by either charge directed or charge remote fragmentation pathways have been examined by molecular orbital calculations. Potential transition state structures were initially examined at the PM3 level of theory, using the simplest model for a methionine sulfoxide containing peptide, N-acetyl-methionine sulfoxide-N-methyl amide. Low energy conformers were then re-optimized at the B3LYP/6-31+G** level of theory followed by vibrational frequency analysis to confirm that each structure represented a saddle point on the potential energy surface. Then, intrinsic reaction coordinate runs were performed, Journal of Proteome Research • Vol. 3, No. 4, 2004 755

research articles

Reid et al.

Figure 1. Quadrupole ion trap tandem mass spectrometry of deuterated GAILM(ox)GAILR. (A) CID MS/MS product ion spectrum of the fully deuterated singly charged ion. (B) CID MS/MS product ion spectrum of the fully deuterated doubly charged ion. (C) Zoom scan spectrum of the methionine sulfoxide side chain cleavage product ion from (A). (D) Zoom scan spectrum of the methionine sulfoxide side chain cleavage product ion from (B).

followed by geometry optimizations at the same level of theory to locate the appropriate precursor and product ion structures associated with each transition state. When comparing the relative transition state barriers associated with two different fragmentation pathways (i.e., methionine sulfoxide side chain cleavage versus cleavage C-terminal to aspartic acid), further conformational searches were carried out to locate the lowest energy precursor ion structure in each case. The total energies, zero point vibrational energies, and relative energies computed for all structures are shown in Table 3. The optimized B3LYP/6-31+G** transition state structure for CH3SOH loss from protonated N-acetyl-methionine sulfoxideN-methyl amide (where the proton is located on the N-terminal amide carbonyl oxygen) via a charge remote cis-1,2 elimination mechanism (equation 3) is shown in Figure 4A [TS1]. Similarly, the optimized B3LYP/6-31+G** transition state structures for CH3SOH loss from protonated N-acetyl-methionine sulfoxideN-methyl amide (where the proton is located on the sulfoxide side chain) via charge directed neighboring group participation reactions involving either the N-terminal amide carbonyl oxygen (eq 4) or the C-terminal amide carbonyl oxygen (equation 5), are shown in Figure 4, parts B and C ([TS2] and [TS3], respectively). The activation energy (B3LYP/6-31+G** + ZPVE) associated with the transition state barrier for the cis1,2 elimination reaction transition state barrier was found to be +25.4 kcal mol-1 relative to the lowest energy conformer found for the protonated precursor ion structure [A]. The activation energies (B3LYP/6-31+G** + ZPVE) for the transition state barriers from the neighboring group participation reactions were found to be +22.6 kcal mol-1 and +22.4 kcal mol-1 respectively, relative to structure [A]. 756


The results of these theoretical calculations are consistent with results from a previous experimental study to examine the loss of CH3SOH from a simple model protonated peptide ion, N-acetyl-methionine sulfoxide-O-methyl ester.52 However, these theoretical results, as well as the previously observed experimental data, are seemingly in conflict with the experimental data obtained in the study presented here using the “tryptic” methionine sulfoxide containing peptide ions, where the charge remote mechanism was demonstrated to be experimentally preferred over the charge directed processes. However, it is important to recognize that in the “tryptic” peptide ions examined here, under the nonmobile proton conditions where enhanced methionine sulfoxide side chain cleavage is observed, the limiting step to fragmentation occurring via charge directed pathways may not be the barrier to the fragmentation reaction itself, but the barrier to proton transfer from the thermodynamically preferred initial site of protonation, i.e., the strongly basic amino acid side chains of arginine or lysine, to the site of cleavage. Given the high proton affinities of these amino acids, this is expected to involve a relatively high energy barrier, such that the charge remote side chain cleavage mechanism becomes the energetically favored process for dissociation. The transition state energy calculated for the charge remote methionine sulfoxide side chain cleavage process has also been compared against that for enhanced C-terminal aspartic acid cleavage, which as mentioned above, has been previously demonstrated to have a dominant effect on the appearance of the product ion spectra of peptide ions under nonmobile proton conditions.17 A number of different pathways have been proposed for enhanced Asp-Xaa cleavage, involving either salt

research articles


Figure 2. Quadrupole ion trap tandem mass spectrometry of deuterated GAILM(ox)GAILK. (A) CID MS/MS product ion spectrum of the fully deuterated singly charged ion. (B) CID MS/MS product ion spectrum of the fully deuterated doubly charged ion. (C) Zoom scan spectrum of the methionine sulfoxide side chain cleavage product ion from (A). (D) Zoom scan spectrum of the methionine sulfoxide side chain cleavage product ion from (B). Table 3. Total Energies (ETotal), Zero Point Vibrational Energies (ZPVE) and Relative Energies (Erel) Computed for the Precursor Ions, Transition States and Product Ion Structures Associated with Each Reaction Pathway at the B3LYP/6-31+G** Level of Theory structure

ETOTAL (Hartree)

ZPVE (kcal mol-1)

Erel (kcal mol-1)a

[A] [TS1] [B] [C] [TS2] [D] [E] [TS3] [F] [G] [H] [TS4] [I] [J] [TS5] [K]

-1048.2833812 -1048.2350959 -1048.2777838 -1048.2824633 -1048.2465463 -1048.2722757 -1048.2639306 -1048.2457310 -1048.2847184 -684.4823595 -684.4635277 -684.4102807 -684.4505018 -1047.9073071 -1047.8602621 -1047.8962973

163.13273 158.21812 160.75893 163.05306 162.66512 163.76331 162.53902 161.91345 163.22623 127.01196 126.28185 125.53046 124.85342 155.76369 151.37609 153.15240

0 +25.4b +1.1b +0.5b +22.6b +7.6b +11.6b +22.4b -0.7b 0 +11.1c +43.7c +17.8c 0 +25.1d +4.3d

a E b Energy rel ) total energy + zero point vibrational energy (ZPVE). relative to structure [A]. c Energy relative to structure [G]. d Energy relative to structure [J].

Figure 3. Quadrupole ion trap energy resolved CID MS/MS breakdown curves of (A) the [M+H]+ and [M+2H]2+ ions of GAILMGAILR and GAILM(ox)GAILR, and (B) the [M+H]+ and [M+2H]2+ ions of GAILMGAILK and GAILM(ox)GAILK.

bridged or charge remote mechanisms.18-25 Bailey et. al. have recently demonstrated, using a series of aspartic acid and arginine containing model peptides, that both mechanisms may be operating within a given peptide ion, depending on

the presence and location of the arginine residue relative to the aspartic acid residue.24 In the work presented here however, we have only considered the transition state barrier associated with the charge remote mechanism. Recently, Paizs et. al.25 have calculated transition state structures for the charge remote loss of NH3 from the protonated model peptide Arg-Asp-amide. From this study, it was predicted that the barriers for this process range in height from 39.9 to 56.9 kcal mol-1, depending Journal of Proteome Research • Vol. 3, No. 4, 2004 757

research articles

Figure 4. Optimized precursor, transition state and product ion structures (at the B3LYP/6-31+G** + ZPVE level of theory) for the loss of methane sulfenic acid from protonated N-acetylmethionine sulfoxide-N-methyl amide by (A) charge remote cis1,2 elimination, (B) charge directed nucleophilic attack from the N-terminal amide carbonyl oxygen, or (C) charge directed nucleophilic attack from the C-terminal amide carbonyl oxygen. The energies (in kcal mol-1) shown for each structure are relative to structure [A].

on the extent of “internal solvation” of the transition state by the neighboring guanidino side chain of the arginine residue. Here, to exclude the possibility of localized charge solvation effects having an artifactual influence on the computed barrier height, i.e., to ensure that the fragmentation was modeled as a truly ‘charge remote’ process, we have calculated the transition state barrier for the charge remote Asp-Xaa amide bond cleavage using a neutral model peptide system, N-acetylaspartic acid-N-methyl amide. The optimized transition state structure is shown in Figure 5A [TS4]. The barrier associated with this transition state was found to be +43.7 kcal mol-1, relative to the lowest energy conformer found for the protonated precursor, structure [G], which is consistent with that previously reported by Paizs. To allow a direct comparison of the energetics computed for this charge remote Asp-Xaa fragmentation process with that for methionine side chain cleavage, an optimized transition state structure for the loss of CH3SOH from neutral N-acetylmethionine sulfoxide-N-methyl amide via the charge remote mechanism was also calculated (Figure 5B [TS5]). It can be seen from Figure 5B that the transition state barrier for this species (+25.1 kcal mol-1 relative to the neutral precursor structure [J]) was very similar to that calculated for the charge remote cis-1,2 elimination reaction from the protonated methionine sulfoxide containing model peptide shown in Figure 4A, and only approximately half that calculated for Asp-Xaa cleavage (Figure 5A). Interestingly, this difference in energies between methionine sulfoxide side chain cleavage and Asp-Xaa cleavage transition states is entirely consistent with the results obtained 758


Reid et al.

Figure 5. Optimized precursor, transition state and product ion structures (at the B3LYP/6-31+G** + ZPVE level of theory) for (A) amide bond fragmentation of neutral N-acetyl-aspartic acidN-methyl amide via a charge remote cis-1,2 elimination reaction mechanism, and (B) the loss of methane sulfenic acid from neutral N-acetyl-methionine sulfoxide-N-methyl amide via a charge remote cis-1,2 elimination reaction mechanism. The energies (in kcal mol-1) shown for each structure in (A) and (B) are relative to structures [G] and [J], respectively.

from statistical analysis of the peptide MS/MS database (Table 1), where methionine sulfoxide cleavage was found to be statistically favored approximately 2-fold over Asp-Xaa cleavage.

Conclusions The results of the study performed here to examine the factors and mechanisms influencing the appearance and abundances of product ion formed by low energy CID of peptides containing methionine sulfoxide, indicate that the formation of enhanced product ions corresponding to the neutral loss of methane sulfenic acid from the side chain of methionine sulfoxide residues occurs under conditions of low proton mobility via a “charge remote” fragmentation process. The abundance of these product ions were found to be statistically at least 2-fold greater than those resulting from the cleavage C-terminal to aspartic acid, which had previously been reported as the dominant fragmentation processes observed under these conditions. This observation was further supported by the results of molecular orbital calculations to predict the transition state barriers associated with these competing reaction pathways. These results clearly indicate that a combined statistical and mechanistic approach to understanding peptide ion fragmentation behavior provides a range of complementary information not available to each method alone. Incorporation of the results arising from this study, as well as those from further studies aimed at systematic analysis of the fragmentation behavior of other common post-translational modifications, will enable the development of more comprehensive algorithms for automated high throughput protein identification and characterization

research articles


based on database analysis of peptide MS/MS data derived from posttranslationally modified proteins.

Acknowledgment. G.E.R. acknowledges the award of an Australian Research Council (ARC) postdoctoral fellowship. References (1) Aebersold, R.; Goodlett, D. R. Chem. Rev. 2001, 101, 269-295. (2) O’Hair, R. A. J. J. Mass Spectrom. 2000, 35, 1377-1381. (3) Clauser, K. R.; Hall, S. C.; Smith, D. M.; Webb, J. W.; Andrews, L. E.; Tran, H. M.; Epstein, L. B; Burlingame, A. L. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 5072-5076. (4) Simpson, R. J.; Connolly, L. M.; Eddes, J. S.; Pereira, J. J.; Moritz, R. L; Reid, G. E. Electrophoresis 2000, 21, 1707-1732. (5) Qin, J; Chait, B. T. Anal. Chem. 1997, 69, 4002-4009. (6) Reid, G. E.; Rasmussen, R. K.; Dorow, D. S; Simpson, R. J. Electrophoresis 1998, 19, 946-955. (7) Swiderek, K. M.; Davis, M. T.; Lee, T. D. Electrophoresis 1998, 19, 989-997. (8) Zugaro, L. M.; Reid, G. E.; Ji, H.; Eddes, J. S.; Murphy, A. C.; Burgess, A. W; Simpson, R. J. Electrophoresis 1998, 19, 867-876. (9) Annan, R. S; Carr, S. A. Anal. Chem. 1996, 68, 3413-3421. (10) DeGnore, J. P.; Qin, J. J. Am. Soc. Mass Spectrom. 1998, 9, 11751188. (11) Hunt, D. F.; Yates, J. R.; Shabanowitz, J.; Winston, S; Hauer, C. R. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 6233-6237. (12) Papayannopoulos, I. A. Mass. Spectrom. Rev. 1995, 14, 49-73. (13) Eng, J. K.; McCormack, A. L; Yates, J. R. J. Am. Soc. Mass Spectrom. 1994, 5, 976-989. (14) Yates, J. R.; Eng, J. K; McCormack, A. L. Anal. Chem. 1995, 67, 3202-3210. (15) Yates, J. R.; Eng, J. K.; McCormack, A. L; Schieltz, D. Anal. Chem. 1995, 67, 1426-1436. (16) Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S. Electrophoresis 1999, 20, 3551-3567. (17) Kapp, E. A.; Schu ¨ tz, F.; Reid, G. E.; Eddes, J. E.; Moritz, R. L.; O’Hair, R. A. J.; Speed, T. P.; Simpson, R. J. Anal. Chem. 2003, 75, 6251-6264. (18) Yu, W.; Vath, J. E.; Huberty, M. C; Martin, S. A. Anal.Chem. 1993, 65, 3015-3023. (19) Summerfield, S. G.; Whiting, A; Gaskell, S. J. Int. J. Mass Spectrom. Ion. Proc. 1997, 162, 149-161. (20) Tsaprailis, G.; Nair, H.; Somogyi, A.; Wysocki, V. H.; Zhong, W.; Futrell, J. H.; Summerfield, S. G.; Gaskell, S. J. J. Am. Chem. Soc. 1999, 121, 5142-5154. (21) Tsaprailis, G.; Somogyi, A.; Nikolaev, E. N; Wysocki, V. H. Int. J. Mass Spectrom. 2000, 195/196, 467-479. (22) Gu, C.; Tsaprailis, G.; Breci, L; Wysocki, V. H. Anal. Chem. 2000, 72, 5804-5813. (23) Huang, Y.; Wysocki, V. H.; Tabb, D. L.; Yates, J. R. Int. J. Mass Spectrom. 2002, 219, 233-244. (24) Bailey, T. H.; Laskin, J.; Futrell, J. H. Int. J. Mass Spectrom. 2003, 222, 313-327. (25) Paizs, B.; Suhai, S.; Hargittai, B.; Hruby, V. J; Somogyi, A. Int. J. Mass Spectrom. 2002, 219, 203-232. (26) Wysocki, V. H.; Tsaprailis, G.; Smith, L. L; Breci, L. A. J. Mass Spectrom. 2000, 35, 1399-1406. (27) Gross, M. Int. J. Mass Spectrom. Ion Proc. 1992, 118/119, 137165. (28) Schutz, F.; Kapp, E. A.; Simpson, R. J.; Speed, T. P. Biochem. Soc. Trans. 2003, 31, 1479-1483. (29) Yalcin, T.; Khouw, C.; Csizamadia, I. G.; Peterson, M. R; Harrison, A. G. J. Am. Soc. Mass Spectrom. 1995, 6, 1164-1174. (30) Yalcin, T.; Csizamadia, I. G.; Peterson, M. R; Harrison, A. G. J. Am. Soc. Mass Spectrom. 1996, 7, 233-242. (31) Ambipathy, K.; Yalcin, T.; Leung, H.-W.; Harrison, A. G. J. Mass Spectrom. 1997, 32, 209-215. (32) Lioe, H.; O’Hair, R. A. J; Reid, G. E. Rapid Commun. Mass Spectrom. 2004, 18, 978-988. (33) Lioe, H.; O’Hair, R. A. J; Reid, G. E. J. Am. Soc. Mass Spectrom. 2004, 15, 65-76. (34) Reid, G. E.; Simpson, R. J; O’Hair, R. A. J. J. Am. Soc. Mass Spectrom. 2000, 11, 1047-1060. (35) Reid, G. E.; Simpson, R. J; O’Hair, R. A. J. Int. J. Mass Spectrom. 1999, 190/191, 209-230. (36) O’Hair, R. A. J.; Styles, M. S; Reid, G. E. J. Am. Soc. Mass Spectrom. 1998, 9, 1275-1284.

(37) Reid, G. E.; Simpson, R. J; O’Hair, R. A. J. J. Am. Soc. Mass Spectrom. 1998, 9, 945-956. (38) O’Hair, R. A. J.; Reid, G. E. Rapid Commun. Mass Spectrom. 1998, 12, 999-1002. (39) Reid, G. E.; Simpson, R. J.; O’Hair, R. A. J. J. Am. Soc. Mass Spectrom. 2000, 11, 1047-1060. (40) Ballard, K. D.; Gaskell, S. J. Int. J. Mass Spectrom. 1991, 111, 173189. (41) Thorne, G. C.; Ballard, K. D.; Gaskell, S. J. J Am. Soc. Mass Spectrom. 1990, 1, 249-257. (42) Yague, J.; Paradela, A.; Ramos, M.; Ogueta, S.; Marina, A.; Barahona, F.; Lopez de Castro, J. A.; Vazquez, J. Anal. Chem. 2003, 75, 1524-1535. (43) Farrugia, J. M.; O’Hair, R. A. J. Int. J. Mass Spectrom. 2003, 222, 229-242. (44) Tabb, D. L.; Smith, L. L.; Breci, L.; Wysocki, V. H.; Lin, D. Y.; Yates, J. R. Anal. Chem. 2003, 75, 1155-1163. (45) Stadtman, E. R.; Moskovitz, J.; Levine, R. L. Antioxid. Redox Signaling 2003, 5, 577-582. (46) Hoshi, T.; Heinemann, S. J. Physiol. 2001, 531, 1-11. (47) Levine, R. L.; Moskovitz, J.; Stadtman, E. R. Int. Union Biochem. Mol. Biol. 2000, 50, 301-307. (48) Stadtman, E. R.; Berlett, B. S. Drug Metabolism Rev. 1998, 30, 225-43. (49) Lagerwerf, F. M.; van de Weert, M.; Heerma, W.; Haverkamp, J. Rapid Commun. Mass Spectrom. 1996, 10, 1905-1910. (50) Jiang, X.; Smith, J. B.; Abraham, E. C. J. Mass Spectrom. 1996, 31, 1309-1310. (51) Betancourt, L.; Takao, T.; Gonzalez, J.; Reyes, O.; Besada, V.; Padron, G.; Shimonishi, Y. Rapid Commun. Mass Spectrom. 1999, 13, 1075-1076. (52) O’Hair, R. A. J.; Reid, G. E. Eur. Mass Spectrom. 1999, 5, 325334. (53) Eddes, J. S.; Kapp, E. A.; Frecklington, D. F.; Connolly, L. M.; Layton, M. J.; Moritz, R. L.; Simpson, R. J. Proteomics 2002, 2, 1097-1103. (54) Hehre, W. J.; Pople, J. A.; Radom, L. Ab Initio Molecular Orbital Theory; Wiley: New York, 1986. (55) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; HeadGordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.7; Gaussian, Inc.: Pittsburgh, PA, 1998. (56) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, K. N.; Kudin, J. C.; Burant, J. M.; Millam, S. S.; Iyengar, J.; Tomasi, V.; Barone, B.; Mennucci, M.; Cossi, G.; Scalmani, N.; Rega, G. A.; Petersson, H.; Nakatsuji, M.; Hada, M.; Ehara, K.; Toyota, R.; Fukuda, J.; Hasegawa, M.; Ishida, T.; Nakajima, Y.; Honda, O.; Kitao, H.; Nakai, M.; Klene, X.; Li, J. E.; Knox, H. P.; Hratchian, J. B.; Cross, C.; Adamo, J.; Jaramillo, R.; Gomperts, R. E.; Stratmann, O.; Yazyev, A. J.; Austin, R.; Cammi, C.; Pomelli, J. W.; Ochterski, P. Y.; Ayala, K.; Morokuma, G. A.; Voth, P.; Salvador, J. J.; Dannenberg, V. G.; Zakrzewski, S.; Dapprich, A. D.; Daniels, M. C.; Strain, O.; Farkas, D. K.; Malick, A. D.; Rabuck, K.; Raghavachari, J. B.; Foresman, J. V.; Ortiz, Q.; Cui, A. G.; Baboul, S.; Clifford, J.; Cioslowski, B. B.; Stefanov, G.; Liu, A.; Liashenko, P.; Piskorz, I.; Komaromi, R. L.; Martin, D. J.; Fox, T.; Keith, M. A.; Al-Laham, C. Y.; Peng, A.; Nanayakkara, M.; Challacombe, P. M. W.; Gill, B.; Johnson, W.; Chen, M. W.; Wong, C.; Gonzalez, and Pople, J. A. Gaussian 03, Gaussian, Inc., Pittsburgh PA, 2003. (57) Lioe, H.; O’Hair, R. A. J.; Gronert, S; Reid, G. E. Manuscript in preparation. (58) Cooks, R. G.; Kruger, T. L. J. Am. Chem. Soc. 1977, 99, 1279-1281. (59) McLuckey, S. A.; Cameron D.; Cooks, R. G. J. Am. Chem. Soc. 1981, 103, 1313-1317. (60) Armentrout, P. B. J. Am. Soc. Mass Spectrom. 2000, 11, 371-379.

PR0499646

Journal of Proteome Research • Vol. 3, No. 4, 2004 759

Statistical and Mechanistic Approaches to ... - ACS Publications

Recommend Documents