Volume 6 Issue 10 - December 12, 2008
Dimethyl Isotope-Coded Affinity Selection (DICAS) for the Analysis of Free and Blocked N-termini of Proteins Using LC-MS/MS
Po-Tsun Shen, Jue-Liang Hsu and Shu-Hui Chen*

Department of Chemistry, College of Sciences, National Cheng Kung University
shchen@mail.ncku.edu.tw

Anal. Chem. 79, 9520-9530, 2007

Font Normal   Font Enlarge
Analysis of protein N-termini is of great importance in both biological and methodological areas.  In the biological area, analysis on N-terminal sequences of proteins could provide many insights such as the proteolytic enzymes and modifications involved in protein synthesis process to assist in better understanding related mechanisms.  For example, during the protein synthesis in many eukaryotes, the initiator methionine of proteins is hydrolyzed and an acetyl group is added to the new N-terminal amino acid.1-3  Proteins that are membrane bound or are destined for excretion contain a signal sequence in their N-termini and it will be removed during the co-translational process.  Moreover, protein N-terminal processing steps are critical for regulating protein turnover in the cell.4,5  In methodological areas, analysis of protein N-termini is termed “positional proteomics”6 and it represents a selective strategy that simplifies the complexity of a proteome by selecting positionally defined peptides that could yield substantial information in protein identification strategies.  Since the two positional locations within every protein are the extreme ends- the N-terminal and C-terminal peptides, analysis of protein N-termini anchors the peptides at a precise location within the parent protein and thus greatly increase the confidences for protein identification.6

Many methods have been reported for selective recovery of N-terminal peptides, including traditional Edman degradation and targeted chemical derivatizations designed to distinguish the N-terminal peptides and internal peptides.  For example, acetylation of free amino groups is coupled with 2,4,6-trinitrobenzenesulfonic acid to cause a strong hydrophobic shift and segregate from the unlabeled N-terminal peptides7 or coupled with biotin-avidin affinity selection to isolate the acetylated N-terminal peptides6; guanidination of ε-amino groups of lysines with o-methylisourea is coupled with biotinylcysteic acid reaction for affinity selection with the -amino group of the protein.8  Isotope-coded N-terminal protein sulphonation has also been reported for simultaneous protein identification and quantitation.9  So far, these reported methods are primarily designed to identify the free N-termini sequence of proteins.  In this report, we aim to analyze both the free and blocked N-termini of proteins simultaneously and demonstrate its potential usefulness in analyzing complex samples. 

A general scheme depicting the procedure for analyzing N-terminal peptides of proteins is depicted in Figure 1.  At the protein level, the disulfide bonds were reduced by DTT and all cysteines were S-alkylated by iodoacetamide.  All primary amines including lysine residues and the N-termini of proteins were reductively aminated by formaldehyde-d2 and sodium cyanoborohydride to produce dimethyl groups or monomethylated proline N-terminus.  The labeled proteins were then digested by trypsin to result in Arg-C like digests.  Internal peptides were then trapped by the solid support (POROS-AL) that contains aldehyde functionalities through the formation of covalent bonds with free amino groups generated by digestion.  The flow-through fractions that contained in-vivo blocked or in-vitro formaldehyde-d2 labeled N-terminal peptides were subjective to sequence analysis by nanoLC-Q-TOF MS/MS.  Stable isotope-based dimethyl labeling strategy has been previously reported by our lab for quantitative proteomics.  This labeling strategy is simple with high reaction yield (near 100%).  More interestingly, this labeling strategy provides a1 ion signal enhancement, which is useful for confirming N-terminal amino acids and modifications.  We would like to take this advantage in developing a MS-based strategy for analyzing protein N-termini.  The method is termed Dimethyl Isotope Coded Affinity Selection (DICAS).
Figure 1. Schematics of the DICAS protocol for selecting protein N-termini

The efficiency of dimethyl labeling at the protein level was first examined using myoglobin as the model.  Dimethyl labeling at the peptide level was reported to be complete with a product yield near 100%.  At the protein level, several factors such as the exposure of the labeling site and the steric hindrance could interfere with the reaction.  To minimize these effects, proteins were denatured and alkylated before the labeling.  Based on the sequence of myoglobin, there are 20 expected labeling sites including 19 lysine residues and the free N-terminus of the protein.  Each labeling site will give a mass difference of 32.0564 Da and therefore, a mass difference around 640 Da is expected if all 20 sites are labeled.  The multiple charge envelopes of both labeled and unlabeled myoglobin were shown in Figure 2.  After deconvolution, only single peak with a molecular weight of 16974.14 Da and 17613.21 Da was yielded for the unlabelled and labelled myoglobin, respectively.  Moreover, the mass difference between the labeled and unlabeled proteins was around 639.07 Da, which is close to the expected value based on the calculation.  Therefore, dimethyl labeling at the protein level was confirmed to be near complete. 

Figure 2. The efficiency of dimethyl labeling for intact myoglobin.  The multiple charge envelop of (A) myoglobin and (B) formaldehyde-d2 labeled myoglobin with their de-convoluted spectra in (C) and (D), respectively.  After deconvolution, only single peak with a molecular weight of 16974.14 Da and 17613.21 Da was yielded for the unlabelled and labelled myoglobin, respectively.  A mass difference of 639.07 Da, which corresponds to all 20 expected sites of myoglobin, was clearly obtained and it indicates a complete labeling.

To confirm the effectiveness of the proposed protocol, a standard protein mixture composed of α-lactalbumin, myoglobin, and hemoglobin was used.  The standard mixture was reduced by DTT, S-alkylated by IAM, and then reductively aminated by formaldehyde-d2 and sodium cyanoborohydride.  After trypsin digestion, the N-terminal peptides were isolated from the flow through fraction and then analyzed by MS.  It was confirmed that the signal intensity of the internal peptides was greatly reduced after affinity removal.  On the contrary, the signal of the N-terminal peptides was found to be enhanced after the isolation and it allowed the MS/MS spectra of N-terminal peptides to be successfully assigned.  The N-termini of α-lactalbumin was assigned to be Glu (E) based on the a1 ion shown in Figure 3, which was the 20th residue in the translated protein.  According to the Swissprot protein database, the sequence (MMSFVSLLLVGILFHATQA) from the first residue to 19th residue was destined to the signal peptide and therefore, the real N-termini of the secreted α-lactalbumin begins with the 20th residue, Glu (E).  Our result is consistent with the report. 
Figure 3 MS/MS spectrum for the N terminal peptide (*EQLTKCEVFR) of α-lactalbumin with 98/53 (protein/peptide) scores. The enhanced a1 ion (m/z 134.12) indicates a free glutamic acid as its N-termini.  This amino acid corresponds to the 20th amino acid ofα-lactalbumin in SwissProt after the removal of a signal sequence (1st-19th amino acid: MMSFVSLLLVGILFHATQA).

We further applied the strategy to the analysis of protein N-termini in MCF-7 cells.  As a proof of the concept, we used the un-fractionated total lysate of MCF-7 cells for the analysis.  A total of 28 proteins were identified from the analysis, among which 23 proteins were identified by single N-terminal peptides, indicating a near 82% of high isolation efficiency.  There were 5 proteins identified from 6 internal peptides and two of them were keratin proteins which were likely to be introduced during the sample handling.  Moreover, 14 out of 23 N-terminal proteins were identified to possess the methionine removal and N-terminal acetylation.  It is notable that there is no detectable a1 ion in the MS/MS spectrum for the N terminal peptide (AcSQAEFE*KAAEEVR, m/z 784.41, 2+) of Acyl-CoA-binding protein, indicating a lack of dimethyl labeling possibly due to the blockage of the N-termini.  Instead, an acetylated serine N-terminus could be readily deduced from b and y ions with a high score.  We further identified a novel signal sequence (1st-32th aminoacid) in the termini of profilin, and the 33th aminoacid, alanine, was identified to be the real N-termini of profilin.  Profilin is a ubiquitous eukaryotic protein that serves as a link between the phosphatidyl inositol cycle and actin polymerisation, and hence it acts as an essential component in the signaling pathway.  Sequence similarity between profilins from different species is low.  However, the N-terminal region, which is thought to be involved in actin binding, is relatively well conserved.  Therefore, the removal of N-terminal signal sequence of profilin is likely to affect the process of actin binding and polymerisation as well as cytoskeletal rearrangement.
< previousnext >
Copyright National Cheng Kung University