A Ubiquitous but Overlooked Side Reaction in Peptide Dimethyl

3 hours ago - Download Citation · Email a Colleague · Order Reprints · Rights & Permissions · Citation Alerts. Add to ACS ChemWorx. SciFinder Subscrib...
0 downloads 0 Views 367KB Size
Subscriber access provided by University of Sunderland

Article

A Ubiquitous but Overlooked Side Reaction in Peptide Dimethyl-labeling. Rijing Liao, Yuan Gao, Ming Chen, Lulu Li, and Xuye Hu Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b03570 • Publication Date (Web): 29 Oct 2018 Downloaded from http://pubs.acs.org on October 30, 2018

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

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

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

Analytical Chemistry

A Ubiquitous but Overlooked Side Reaction in Peptide Dimethyllabeling Rijing Liao,*,† Yuan Gao, ‡ Ming Chen, ⊥ Lulu Li,† and Xuye Hu§ Institute of Precision Medicine, The Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China †



380 Xinlongcheng, Changping District, Beijing, China 102206

State Key Laboratory of Bioorganic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 345 Lingling Road, Shanghai 200032, China § Shanghai Clinical Center, Chinese Academy of Sciences, Shanghai 200031, China ⊥

ABSTRACT: Reductive dimethylation using formaldehyde and NaBH3CN to label peptides or proteins on their N-termini and lysine residues is one of the most widely used labeling methods in quantitative proteomics field. In this study, we characterized a ubiquitous side reaction in dimethylation labeling, causing a mass increment of 26 Da on the N-termini of peptides. It can extensively occur on most peptides, which significantly compromises data quality in terms of sensitivity, dynamic range and peptide/protein identification rates. Nevertheless, this side reaction was so-far overlooked largely due to the current database search algorithms limited for the detection of unknown modifications. In order to illustrate the chemical nature of this side reaction, 1D and 2D nuclear magnetic resonance (NMR) was performed to elucidate the exact structure of the modification formed through this side reaction, revealing that the side reaction produced a N-methyl-4-imidazolidinone moiety between the first two residues of the undesirably labeled peptides. Based on the mechanism proposed for the side reaction, we optimized the dimethyl-labeling reaction conditions. Compared to the current typical labeling method, our approach can dramatically suppress the side reaction at both standard protein and proteome levels. As a result, with our optimal labeling method, peptide and protein identification rates were significantly increased compared to those with traditional labeling method.

INTRODUCTION The advent of mass spectrometry (MS) in combination with liquid chromatography (LC) has enabled proteomics to identify and quantify thousands of proteins from complex biological samples. Nevertheless, the accurate quantification of proteins across multiple biological samples is not without difficulty. The signal intensities or extracted ion current (XIC) peak areas of peptides are commonly used when comparing their abundances across multiple samples. Unfortunately, the MS signal intensity of a given peptide is not only determined by its abundance, but also prone to impacts of several other factors (e.g., elution time and matrix effects). Thus, the variations between separate LC fractionation and LC-MS analysis are inevitable and will compromise the accuracy of protein relative quantification between different samples. To minimize the variation, stable isotope-labeling strategies have been developed to enable processing and analyzing multiple samples under the same chromatography and mass spectrometry conditions. Samples to be compared are differentially labeled with isotopic tags (a set of isotopic variants), and then mixed before further processing and LC-MS/MS analysis. Peptides/proteins from different samples can be distinguished based on their isotopic tags, which are isotopically distinct in MS or MS/MS spectra. The benefit is that, these isotopic tags are chemically identical, thus the cognate peptides from different samples will co-elute in chromatographic separation and have the same electrospray ionization efficiency in MS analysis, which improves the accuracy of quantification when comparing to the label-free approaches.

Among reported stable isotopic labeling methods, reductive dimethylation1-6 of peptides with formaldehyde in the present of sodium cyanoborohydride (NaBH3CN) is, in terms of procedure, straightforward, fast and undemanding. Furthermore, this method is quite budget friendly in comparison with other isotope-labeling methods such as SILAC (stable isotope labeling by amino acid in cell culture)7 and iTRAQ (isobaric tag for relative and absolute quantification)8. During dimethyl labeling, the primary amine-groups of peptides, including the N-terminal amines and the -amines of lysine residues, can react with formaldehyde to form the imine (also called Schiffbase) intermediates. The imine intermediates will be further reduced to give the N-methyl group, in the presence of reducing agent such as NaBH3CN. However, the imine intermediates are also reactive with the side-chains of many amino acid residues (e.g. arginine, histidine and tyrosine), resulting in intra- or interpeptide cross-links.9 Undoubtedly, these side reactions will compromise the quality of analysis in terms of sensitivity, dynamic range, and peptide/protein identification rates. Side reactions reduce the yield of the desirable products, therefore reducing their chance of detection. Moreover, the byproducts produced by side reactions increase the sample complexity, which further reduce the probability of low-abundance peptides being selected for MS/MS fragmentation in the data dependent acquisition (DDA) mode. However, side reactions in labeling are often overlooked by proteomic researchers and consequently the chemical nature of their products remains unclear. This is largely due to the current database search algorithms that cannot identify the side products without knowing the exact mass and positions of the modifications caused by the side reactions.

ACS Paragon Plus Environment

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

In this work, we characterized a side reaction, which could occur significantly in reductive dimethylation labeling of peptides with formaldehyde and NaBH3CN. We first observed this side reaction on a synthetic peptide, and further analysis of complex biological samples revealed that it can extensively occur on most peptides, resulting in a 26 Da mass increment in the N-termini of peptides. The exact chemical structure of the modification moiety caused by side reaction was elucidated by 1D and 2D nuclear magnetic resonance (NMR, including 1H NMR, 13C NMR, COSY, HSQC, and HMBC), revealing that a N-methyl-4-imidazolidone moiety was formed between the first and second residues of peptides in the side reaction. The illustrated structure suggests that the side reaction undergoes an imine intermediate, which is further converted to the imidazolidone via an intra-peptide crosslinking with the adjacent amide group. Based on this proposed mechanism, we optimized the labeling conditions to suppress the side reactions in dimethyl-labeling. Using the optimized labeling conditions, we can suppress this side reactions and increase the peptide identification rates. We demonstrated that about 15% more peptides and 10% more proteins were identified in the same the Hela tryptic digests with the optimal labeling method compared to that with the traditional labeling method.

MATERIALS AND METHODS Chemicals and Reagents. Chemicals were purchased from Sigma-Aldrich unless stated elsewhere. The double deionized water was produced by Milli-Q gradient A10 system (Millipore, Bedford, MA). The standard peptides were purchased from GL Biochem Ltd (Shanghai, China). The Bovine Serum Albumin (BSA) standard protein was purchased from Beyotime Ltd (Shanghai, China). Mammalian Cell culture and Protein Preparation. HeLa 229 cells were cultured as the previously described.10 The cells were washed once in PBS buffer, and then were sonicated in 6 M urea/2 M thiourea on ice until the lysate appeared clear. The proteins were reduced with 10 mM dithiothreitol (DTT) at 37 °C and then alkylated by 25 mM iodoacetamide at room temperature for 30 min in the dark. The protein solution was mixed with 5 times of ice-cold acetone and incubated at -20 °C overnight. The precipitate was collected by centrifugation and washed twice with ice-cold acetone. The pellet was dissolved in 1 M urea and 50 mM ammonium bicarbonate (pH 8.0), followed by trypsin digestion at 37 °C overnight. After a desalting step using Sep-PAK C18 (Waters), the tryptic peptides were dried by a SpeedVac and stored at -80 °C until use. Dimethylation Labeling. Two labeling methods were used in this study to compare their labeling performance. All labeling reactions and quenching step should be carried out in a fume hood to protect the researchers. In the traditional method, peptides were dissolved in 100 L of 100 mM triethylammonium bicarbonate (TEAB, pH 8.5). 4 L of 4% (v/v) formaldehyde solution was added into the peptide solution, followed by brief mixing and centrifugation. Immediately, 4 L of 0.6 M NaBH3CN was added into the sample, followed by brief mixing and centrifugation. After incubation at 30 °C for 1

Page 2 of 8

h, the labeling reaction was quenched by adding 16 L of 1% (v/v) ammonia solution. In our optimized method, peptides were dissolved in 100 L of 100 mM NaH2PO4 (pH 6.0). Immediately before use, 4% (v/v) formaldehyde were mixed with 1.2 M NaBH3CN in the ratio of 1:1 (v/v), then the 8 L of the mixture was added into the peptides solution, followed by brief mixing and centrifugation. The labeling reaction was incubated at 30 °C for 1h, and then quenched by adding 16 L of 1% (v/v) ammonia solution. Elucidation of byproduct structure by NMR. 10 mg of the peptide LVNE was dissolved in 10 mL of 100 mM triethylammonium bicarbonate (TEAB, pH 8.5). 400 L of 4% (v/v) formaldehyde solution was added into the peptides solution, followed by the addition of 400 L of 0.6 M NaBH3CN. After incubation at 30 °C for 1 h, the labeling reaction was quenched by adding 1.6 mL of 1% (v/v) ammonia solution. After desalting using a Waters Sep-pak C18 column, the labeling products were further purified by semi-preparative reversed-phase HPLC (Agilent series 1200 system). About 1 mg of the byproduct with a mass increment of 26 Da was obtained. The byproduct was dissolved in 0.5 mL of CD3OD and then subjected to a Bruker 600 MHz NMR for the acquisition of the 1H NMR, 13C NMR, COSY, HSQC, and HMBC data. Nano-RPLC-EThcD-MS/MS. Samples were loaded onto a trap column (75 m i.d., 2 cm, C18, 3 m, 100Å, Dionex) with an online desalting wash, and then separated with a reversed phased column (75 m i.d.,10.2 cm, C18, 3 m, 120Å, New Objective) by an EASY-nLC 1000 (Thermo Scientific) at a flow rate of 300 nL/min. The synthetic peptide samples were separated using a 30 min LC gradient (10-50% B from 0 to 30 min; solvent A, 0.1 % formic acid in H2O; solvent B, 0.1 % formic acid in acetonitrile), BSA tryptic digests were separated using a 1 h LC gradient (6−24% B over 40 min, then 24−40% B from 40 to 47 min, and 40−90% B from 47 to 50 min, continuing at 90% B up to 60 min), and the HeLa samples were separated with a 3 h LC gradient (5−22% B over 115 min, then 22−40% B from 115 to 163 min, and 40−90% B from 163 to 167 min, continuing at 90% B up to 180 min). The nano-ESI voltage was set to 2.0 kV and the capillary temperature was 275 °C. The MS data were acquired on an Orbitrap Fusion mass spectrometer (Thermo Scientific) in a data-dependent mode. Each full-scan MS (m/z 350-2000, resolution of 60 k) was followed with several HCD MS/MS scans (normalized collision energy 30, m/z 150-2000, resolution of 15 k) for the most intense precursor ions within a cycle time of 3 s. The maximum ion injection time (MIT) for MS1 and MS/MS were 50 and 100 ms, and the auto gain control target for MS1 and MS/MS were 2 × 105 and 2 × 105, respectively. The dynamic exclusion time was set as 60 s. Data Analysis. Raw data were viewed in Xcalibur (Thermo Scientific). Proteome Discoverer 2.2 (Thermo Scientific) with the SEQUST engine was used for database search and quantification. The raw data of BSA and HeLa samples were searched against to the bovine Uniprot or human UniProtKB database, respectively. Enzyme specificity was set to trypsin, allowing for up to 2 missed cleavages. Mass tolerance was set

ACS Paragon Plus Environment

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

Analytical Chemistry to 10 ppm for the precursor ions and 0.02 Da for fragment ions. The carbamidomethylation (+57.0215 Da) on cysteine, and dimethylation (light: +28.0313 Da, heavy: +30.0380 Da) on Nterminal amine and lysine were set as static modifications. In order to assess the level of side products in labelling samples, database searches were carried out with the same settings as above except using a fixed +26.0156 Da modification to replace the N-terminal dimethylation. The decoy database searches were also performed in parallel, and peptides less than 1% false discovery rate (FDR) were accepted.

RESULTS AND DISCUSSION We first noted this side reaction on a synthetic peptide DRVYIHPF that was dimethyl-labeled with formaldehyde and NaBH3CN as described previously.3 Analyzing the labeling products with LC-MS, the base peak chromatogram showed two peaks eluted at 13.02 min (m/z 537.7903) and 16.02 min (m/z 536.7827), respectively (Figure 1a). Confirmed by the MS/MS spectra, the major peak at 13.02 min was the desirable product with correct dimethylation at N-terminal amino group (Figure 1b). However, the minor peak at16.02 min was an unknown product 26 Da heavier than the unmodified peptide. Comparing the MS/MS spectra from the correctly dimethyllabeled peptide or the side product, both spectra had the same y-ion series but distinct b-ion series (Figure 1b, c), suggesting that the side product was an adduct of DRVYIHPF with the modification appended at the peptide N-termini. Especially, the y7 fragment ions (m/z 931.51) detected in the both MS/MS spectra further supported the hypothesis that the modification was just attached at the first residue.

spectra of the correctly dimethylated peptide (b) or the side product (c). In order to assess the frequency of this side reaction, we then applied the dimethyl-labeling reaction on BSA tryptic digests following the same procedure. The labeling products were subjected to LC-MS/MS for analysis, and the raw data was processed by Proteome Discoverer software (Materials and Methods). The database search results showed that this side reaction occurred on most BSA tryptic peptides. In the BSA digests, 42 peptides were identified as the correctly dimethyllabeled, among which 31 peptides were also detected as the undesirably labeled. Moreover, there were 4 peptides detected only in the undesirably labeled form. Table 1 showed the 31 peptides identified in both the correct dimethylation and the undesirably modified forms. Noting that the XCorr score of each undesirably labeled peptide was quite comparable to that of the corresponding correctly labeled species, suggesting that the +26 Da modified peptides were not likely to be false positives or artifacts of the search. Interestingly, each +26 Da modified peptide showed longer retention time than the corresponding dimethylated peptide, indicating that this +26 Da modification caused larger increase in peptide hydrophobicity than the dimethylation did. Table 1. BSA tryptic peptides identified in both dimethylation form and the +26 Da modified form through traditional dimethyl labeling method. Peptide sequences HPEYAVSVLLR HLVDEPQNLIK CCTESLVNR SHCIAEVEK DLGEEHFK NECFLSHK ETYGDMADCCEK GACLLPK LGEYGFQNALIVR EACFAVEGPK LFTFHADICTLPDTEK QTALVELLK LVNELTEFAK NECFLSHKDDSPDLPK DAIPENLPPLTADFAEDK TCVADESHAGCEK DDSPDLPK TVMENFVAFVDK YICDNQDTISSK HPYFYAPELLYYANK SLHTLFGDELCK DAFLGSFLYEYSR KQTALVELLK EYEATLEECCAK ECCHGDLLECADDR GLVLIAFSQYLQQCPFDEHVK DDPHACYSTVFDK LKPDPNTLCDEFK CCAADDKEACFAVEGPK RHPEYAVSVLLR DAIPENLPPLTADFAEDKDVCK

Xcorr ratios between dimethyl+26 Da labeled modified two speciesa 2.76 3.77 0.20 3.98 4.59 0.12 2.18 2.58 1.78 2.60 3.00 3.55 2.38 2.53 0.18 2.27 2.37 1.94 4.16 4.21 0.15 2.26 2.25 0.04 3.92 3.85 0.01 3.12 3.02 0.04 4.28 4.18 0.69 3.23 3.09 0.05 3.26 3.11 0.23 3.73 3.58 0.15 3.25 3.07 1.44 3.76 3.53 1.04 2.59 2.31 0.37 4.72 4.37 0.02 3.44 3.08 0.23 3.88 3.47 0.61 4.53 3.90 0.30 3.96 3.29 0.03 3.58 2.73 0.32 4.53 3.60 0.09 5.14 3.98 0.67 4.86 3.34 0.01 4.20 2.51 0.25 5.08 3.32 0.03 5.19 3.38 0.41 5.97 4.07 0.03 5.41 2.92 0.15

ratio = intensity of +26-Da-modified peptides / intensity of dimethyl-peptide a

Figure 1. LC-MS/MS analysis of the dimethyl-labeling products of the synthetic peptide DRVYIHPF. (a) The base peak chromatogram of the labeling products. The MS/MS

The observation that this +26 Da modification could extensively occur on most BSA tryptic peptides motivated us to study it in detail. The fact that this side reaction can happen to the peptides starting by glycine or aliphatic amino acids (Table

ACS Paragon Plus Environment

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

1) indicates that side chains of amino acids are not involved in the side reaction. The 26.0162 Da mass increment of the side product, corresponding to a nominal C2H2 moiety, suggested the increase of 1 degree of unsaturation in side products. In another word, either one ring or one double bonds were introduced to the side products by the side reaction. But, given the fact that the labeling reactions were carried out in the presence of NaBH3CN, which was a reducing reagent that is strong enough to reduce double bonds, a ring structure was more likely to be the one formed in the side products. One possible structure of the side product was N-methyl-4-imidazolidinone (Supporting Information Figure S1a). A previous study conducted by Wim Jiskoot et al,9 has reported that formaldehyde, in the absence of NaBH3CN, could cross-link the N-terminus amine group with the adjacent amide group via a methylene bridge, forming a 4imidazolidinone moiety with a mass increase of 12 Da. In the presence of NaBH3CN, the 4-imidazolidinone would be further converted to N-methyl-4-imidazolidinone. The structure of Nmethyl-4-imidazolidinone exactly fit the mass increase observed in the side product, and it could also explain the longer retention time in the side products than in the dimethylated peptides. However, there were still two hurdles to confidently assign the modification moiety as N-methyl-4-imidazolidinone. Firstly, the reaction conditions reported to yield N-methyl-4imidazolidinone was quite different from traditional dimethyl labelling protocol. In the previous report,9 peptides were first incubated with mere formaldehyde for several hours before the addition of NaBH3CN, allowing the formation of 4imidazolidinone within the first two amino acid residues. In dimethyl-labeling reaction, however, NaBH3CN was immediately added after the addition of formaldehyde. More importantly, the structure of methylene bridge between two residues was inconsistent with the MS/MS result (Figure 1c), which suggested that the modification should located at the first amino acid residue of peptides. In view of the disagreement between 4-imidazolidinone and the MS/MS spectra, we also proposed another possible structure, N-methylaziridine (Supporting Information Figure S1b), for the side product. This structure also introduced a nominal C2H2 moiety to the products. As N-methylaziridine moiety was more hydrophobic than dimethylated amine group, it could also account for the fact that the side products had longer retention time than the corresponding dimethylated products. Furthermore, the structure of N-methylaziridine was consistent with the result of MS/MS spectra, which showed that the modification located at the first amino acid residue. Nevertheless, in spite of the hints provided from retention time, MS and MS/MS spectra, more evidence was required to determine which structure was the right one. To date, nuclear magnetic resonance (NMR) is still the most powerful tool that can provide detailed and decisive information for the structural elucidation of small molecules. In order to elucidate the exact chemical structure of the side product, we decided to determine the structure by high resolution NMR. Identification of the chemical sructure of side product by NMR. In the dimethyl-labeling of BSA tryptic digests, peptide LVNELTEFAK was shown to produce relatively high percentage of the side product. Besides, its first and second amino acid residues have aliphatic side chains, preventing any other side reaction involving the side chains. Therefore, we

Page 4 of 8

decided to use it as a model peptide to elucidate the chemical structure of the side product formed in dimethyl-labeling. In order to reduce the complexity of the NMR spectra, a shortened peptide LVNE was synthesized and used instead of the LVNELTEFAK. The tetrapeptide was labeled with formaldehyde and NaBH3CN at a large scale following the same protocol we used for the dimethyl-labeling of BSA digests. After multiple steps of isolation, about 1 mg of the byproduct was purified and reconstructed in deuterated methanol for the acquisition of the 1H NMR, 13C NMR, COSY, HSQC, and HMBC data (see Materials and Methods for details). The 1H NMR, 13C NMR (Table 2) and HSQC spectra (Supporting information Figure S2a, b, c) indicated that the byproduct had five methyl groups, five methylenes (CH2), six methines (CH), and six carbonyl carbons (C 168.7-174.9). The presence of six methines ruled out the possibility that the byproduct possessed a N-methylaziridine moiety, since the LVNE peptide containing a N-methylaziridine moiety should have 5 methines. The resonances at H 3.02 and C 38.2 were attributed to the N-methyl group. The proton of the N-methyl showed long-range HMBC (Supporting information Figure S2e) correlations to the methine carbon at C 64.2 and methylene carbon at C 65.4, allowing for the assignment of the methine and methylene as the -C of Leu and the methylene in 4imidazolidine, respectively. The assignment of methylene in 4imidazolidine was further supported by the HMBC correlations from H-7 (H 4.58/4.99) to C-6 (C 38.2) and the long-range 1H13C correlation from H-9 ( 4.35) to C-7 ( 65.4). Taken H C together, the chemical structure of the modification moiety in the byproduct was established as an N-methyl-4imidazolidinone moiety between the Leu and Val residues (Figure 2). The key 1H-1H COSY and HMBC correlations in the adduct were also shown in Figure 2.

Table 2. The 1H (600 MHz) and 13C (150 MHz) NMR Data for the adduct of LVNE in CD3OD. C no.a C, type H 1 20.93, CH3 2 24.47, CH 3 21.37, CH3 4 36.59, CH2 Leu 5 64.19, CH 6 38.17, CH3 7 65.37, CH2 8 169.18, C 9 60.96, CH 10 27.46, CH Val 11 18.08, CH3 12 17.17, CH3 13 168.74, C 14 50.20, CH 15 36.27, CH2 Asn 16 172.95, C 17 171.34, C 18 51.62, CH 19 26.57, CH2 Glu 20 29.62, CH2 21 174.08, C 22 173.02, C a The C no. were labeled in Figure 2.

ACS Paragon Plus Environment

1.03, d 2.07, m 1.03, d 1.85, m; 1.69, m 3.89, br 3.02, s 4.58, d; 4.99, d 4.35, d 2.19, m 1.03, d 0.93, d

4.78, dd 2.80, dd; 2.65, dd

4.45, dd 2.21, m; 1.95, m 2.43, t

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

Analytical Chemistry

O

Val Leu

O

11

12 10

4

1 2

5

N

3

9

8

N 7

13

H N

17

15

OH Glu

20 19

1

H-1H COSY

OH

18 14

O

6

Asn O

21

N H O

22

1

H-13C HMBC

O

16

NH2

Figure 2. The key COSY and HMBC correlations of the adduct of the LVNE peptide produced by side reaction.

Optimization of the Dimethyl-Labeling Conditions. Based on the structure elucidation, this ubiquitous side reaction of dimethyl-labeling was established as the intermolecular crosslinking reaction between the N-terminal amine group and the adjacent amide group. The mechanism for this side reaction was proposed and shown in the Scheme 1. In the side reaction, the N-terminal amine groups of peptides first react with formaldehyde and form hydroxymethyl groups. The hydroxymethyl groups will undergo condensation to give the imine groups (Schiff-base), which were also the key intermediates for the desirable dimethylation reactions. In the present of reducing agents, e.g. NaBH3CN, most imine intermediates are reduced quickly to the desirable dimethylated products while a small proportion of imine intermediates undergo the intramolecular nucleophilic addition to form the unwanted side products. Thus, the key factor affects the ratio between the demethylated products and the by-products is the reactivity of the reductant used in the labeling reaction. Obviously, enhancing the reducibility of NaBH3CN would accelerate the reduction of imine and increase the yield of desirable methylated peptides. HN R1

NH2 R1 O

H N

H pep.

fast

OH

O H

HN

H N

R1 O

pep.

O primary

NaBH3CN

H

Me H N

N

-H2O pep.

R1

H N

pep. O imine intermediate NH

slow

N pep. O side product

R1

Scheme 1. The primary reaction and side reaction involved in dimethyl-labeling of peptides with formaldehyde and NaBH3CN.

In most published protocols, dimethyl-labeling was performed in triethyl ammonium bicarbonate (TEAB) buffer and the addition of formaldehyde is prior to that of NaBH3CN.25 Theoretically, adding formaldehyde first might increase the side reactions to some extent because the imine intermediates will form after the addition of formaldehyde, in the absence of NaBH3CN, the imine will be converted to the unwanted side products only. Therefore, we premixed formaldehyde and NaBH3CN immediately before their addition to the labeling solution, allowing the simultaneous addition of both formaldehyde and NaBH3CN. Our experiments indicated that adding the labeling reagents simultaneously did reduce the

peptide spectra matches (PSMs) of the undesirably labeled species when comparing with the traditional method (from 107 of PSMs to 95 of PSMs on average, Figure 3). Thus, in our following experiments, we always premixed formaldehyde and NaBH3CN immediately before their use and added them simultaneously to start the labeling reaction. In order to increase the reducing reaction rate for Schiff-base intermediates, we doubly increased the concentration of NaBH3CN (final concentration 44.4 mM) in TEAB buffer, which was expected to promote the conversion of imine intermediates to methylated products while competitively suppressing the unwanted side reactions. As shown in the Figure 3, increasing NaBH3CN concentration in the labeling solution further reduced the undesirably labeled PSMs (from 95 of PSMs to 84 of PSMs on average, Figure 3). dimethyl-labeled

500 400

352

undesirably labeled 432

420

390

421

441

300 200

107

95

100 0

84 19

TEAB

TEAB premix

TEAB NaBH3CNx2

pH6

9 pH4

15 pH6 NaBH3CNx2

Figure 3. The PSMs numbers of the dimethyl-labeled and the undesirably labeled peptides in BSA tryptic digests labeled under different conditions. The all numbers are the arithmetic means of 3 technical replicates.

In organic synthesis, NaBH3CN is a mild reducing agent often used in acidic reaction solutions rather than basic solution.11, 12 We therefore tried to labeled the BSA digests in a slightly acidic phosphate buffer (100 mM NaH2PO4, pH 6.0). The results indicated that it could dramatically reduce the side products when comparing to the TEAB buffer (from 84 of PSMs to 19 of PSMs on average, Figure 3). Further decreasing the pH value of the phosphate buffer from pH 6.0 to pH 4.0 could slightly reduce the side products. However, the PSMs of the correctly labeled peptides were also decreased (from 432 of PSMs to 421 of PSMs on average, Figure 3). This was likely because the lower pH value also reduced the reaction efficiency of the desirable dimethyl labeling while reducing the side reactions. As our previous experiments showed that doubling the concentration of NaBH3CN could reduce PSMs number of byproducts while increasing the PSMs number of correctly labeled species in TEAB buffer, we doubled the NaBH3CN concentration in the pH6.0 phosphate buffer, allowing further reduction in the PSMs of by-products while increasing the PSMs of correctly labeled peptides (Figure 3). It was worth noting that the labeling efficiency under all the labeling conditions was high, since only negligible PSMs (