|Home | About | Journals | Submit | Contact Us | Français|
Conditions for carrying out chemically targeted identification of peptides containing phosphorylated or glycosylated serine residues have been investigated. Ba(OH)2 was used at ambient temperature to catalyze the β-elimination reaction at 25°C. Nucleophilic addition of 2-aminoethanethiol was performed in both parallel and tandem experiments. The method was demonstrated by the reaction of β-casein tryptic digest phosphopeptides and an O-glycosylated peptide. Contrary to an earlier report by others, the glycopeptide was found to react with essentially the same kinetics as phosphopeptides. Conversion of four phosphoserines in residues 15, 17, 18, and 19 from bovine β-casein N-terminal tryptic phosphopeptides were followed by monitoring the time course of the addition reaction. The chemistry proceeded rapidly at room temperature with a half-reaction time of 15 min. No side-reaction products were observed; however, care was taken to minimize all counter ions that either precipitate barium or neutralize the base. Digestion of the converted peptides with lysine endopeptidase identified all five phosphoserines in the β-casein tryptic digest. Alternatively, preincubation with base followed by nucleophilic addition of the thiol was found to work satisfactorily. The use of the water-soluble hydrochloride of 2-aminoethanethiol allowed β-elimination, nucleophilic addition, and desalting to be carried out on a micro C18 reverse phase pipette tip.
In an earlier communication we reported a combined chemical and enzymatic approach which enabled the identification of phosphorylated and O-glycosylated serine and threonine residues by either short N-terminal sequencing or mass measurement.1 Phosphopeptides and glycopeptides were targeted for site-specific cleavage with a chemistry called chemically targeted identification (CTID).2 The procedure is based on an approach developed by Cole in which a cysteine residue was converted to an analog of lysine and cleaved by a protease.3,4 While the method was shown to work well with both phosphorylated and O-glycosylated serine and threonine peptides containing single modifications,1 questions remained. Complications from serial tandem modifications, incomplete conversion, unwanted side reactions, and local residues that might interfere with proteolysis, posed questions for the method’s usefulness. Recently, a publication appeared cautioning against the use of higher temperatures such as those used in the CTID procedure for β-elimination.5 Anecdotally, we had received reports questioning the robustness and reliability of the chemistry. Taken together, these seemed justifiable reasons to refine conditions for carrying out CTID.
A multiply phosphorylated peptide containing four phosphoserines, three of which occurred in tandem and were flanked by acidic glutamate residues, was deemed a good test for any conditions that might be developed.2,6 The kinetics of the new reaction conditions would be followed and compared with prior results. The same conditions would be applied to an O-glycosylated serine to see whether the new conditions would catalyze deglycosylation.7
Bovine β-casein, S-2-aminoethanethiol hydrochloride (AET), barium hydroxide octahydrate, and 2,5-dihydroxybenzoic acid (DHB) were purchased from Sigma-Aldrich Chemicals (St. Louis, MO). Sequencing grade trypsin (porcine) was purchased from Promega Biochemicals (Madison, WI) and lysine endopeptidase (Lys-C) (Achromobacter lyticus E.C. 18.104.22.168) was obtained from WAKO Chemicals (Richmond, VA). Trifluoroacetic acid (TFA) was from Applied Biosystems (Foster City, CA), and micro C18 ZipTip pipette tips were obtained from Millipore (Bedford, MA). The O-glycosylated peptide Phe-Ala-Ala-(O-GlcNacSer)-Asn-Tyr-Pro-Ala-Leu was the kind gift of Dr. Mona Shagaholi, California Institute of Technology.
Excess barium hydroxide octahydrate [Ba(OH)2], calculated to give a saturated solution at room temperature (126 mg/mL), was added to doubly deionized water and dissolved by gentle swirling.8 The saturated solution was overlaid with argon and tightly stoppered. The stock solution was used for one week, after which it was discarded. Prior to use, aliquots were briefly centrifuged to remove undissolved reagent and precipitates (mostly carbonates). To estimate the reagent’s concentration, equal volumes of saturated supernatant and 0.5 M ammonium bicarbonate were transferred to preweighed micro centrifuge tubes, vortexed, and centrifuged. The pellets were washed once with deionized water and re-centrifuged. Precipitates from triplicate samples were dried in a speedvac centrifuge and their weights recorded.
β-Casein, 2 mg in 20 μL of deionized 8 M urea, was incubated at 37°C for 20 min. The sample was diluted to 2 M urea by adding 60 μL of 0.1 M ammonium bicarbonate, pH 8, that contained 20 μg trypsin. Proteolysis was allowed to proceed overnight at 25°C after which TFA was added to 1% and 2 μL of digest was purified by reversed-phase high-performance liquid chromatography. Fractions containing a 2.061-kDa peptide FQPSEEQQQTEDELQDK and partially purified 3.121-kDa phosphopeptide RELEELNVPGEIVEPSLPSPSPSEESITR were collected and used for the CTID experiments. Peptide concentrations were determined by Edman degradation with a Procise cLC (model 492, Applied Biosystems). Concentration was obtained dividing initial yield by 0.6.
AET was prepared fresh before each experiment as a 0.2 M solution (22.72 mg/mL) in deionized water. The pH 5 solution was used directly, or adjusted to pH 8 with Ba(OH)2. All experiments were conducted at 25°C unless noted otherwise in the legends to the figures (room temperature was measured with a total immersion thermometer). Three different methods were used which differ by the order of addition of reagents and chemicals.
Method 1: Peptides were dried in the vacuum centrifuge and equal volumes of 0.2 M AET and 180 mM Ba(OH)2 were added together.
Method 2: Ba(OH)2 was added in the absence of thiol, and β-elimination was allowed to proceed. Then AET was added and nucleophilic addition was allowed to proceed. At the times indicated in the figures, aliquots were removed and acidified with 10% formic acid (pH 4). The samples were diluted further by mixing 1:10 or 1:250 v/v with DHB matrix and analyzed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). Alternatively, samples were desalted by reversed-phase, and eluted with 3 μL of 45% ACN/0.1% TFA.
Method 3: Peptide was bound to a C18 reversed-phase micropipette tip in 0.005% TFA. The bound peptide was flushed 3 times with 90 mM Ba(OH)2 and the tip was allowed to sit immersed in 5 μL of the base for 30 min. The pipette was flushed 10 times with 200 mM AET which had been pH adjusted with a little Ba(OH)2 to pH 8. The tip was allowed to sit immersed in 5 μL of AET for 60 min. Following the last incubation, the tip was rinsed 10 times with 0.1% TFA and eluted with 1–2 μL of 45% ACN/5% formic acid to the dry DHB matrix spot.
Digestion was carried out in a volume of 20 μL containing the modified peptide, 20 mM Tris-HCl, pH 9.2 and 0.1 μg Lys-C. After overnight digestion at room temperature, the reactions were terminated by adding 1 μL 10% formic acid. Aliquots of 1.0 μL were diluted with 4 μL DHB matrix solution and 0.5 μL was analyzed by MALDI-TOF-MS.
MALDI-TOF-MS was performed with a Voyager-DE STR (Applied Biosystems). The matrix used was DHB (50 mg/mL in 20% ACN, 0.05% TFA). In some experiments, 20 mM ammonium phosphate was added to the matrix to reduce metal complexes and improve spectra.9 A 0.5-μL aliquot of the matrix was spotted to the target and allowed to dry. Thereafter, 0.5–1.0 μL of sample was placed on top of the spot and dried in a stream of 30°C air. Monoisotopic values were reported unless noted otherwise in the legends to the tables and figures. Spectra from the time-course reactions were measured in triplicate, and the monoisotopic peak recorded. The data were smoothed to record average peak heights. Results were expressed as the ratio of the peak height of the product to the sum of the peak heights for reactant and products and plotted as percent versus time. Product ions were confirmed with a QqTOF mass spectrometer (Qstar XL Hybrid, Applied Biosystems), using CAD fragmentation.10
The sequences and calculated mass of the phosphorylated tryptic peptides of bovine β-casein are shown in Table 11.11,12 Four serines located in the first 19 residues (serine 15, 17, 18, 19) of the N-terminal tryptic peptide and a fifth, serine residue 35, have been shown to be phosphorylated.12 The calculated mass of the corresponding 2-aminoethylcysteine product is also shown in Table 11.. The elimination and addition reactions were followed by simultaneously measuring the appearance of newly derivatized phosphopeptides that appear with a net loss of mass 21 (residues shown as “X” in Table 11).). The Ba(OH)2 concentration, estimated gravimetrically, gave triplicate values with a precision of 1% (data not shown). We conservatively estimated a concentration of 0.18 ± 0.01 M in reasonable agreement with published values.8
To test the effect of AET on the conversion, the 2.061-kDa phosphopeptide and the 1.156-kDa glycosylpeptide were incubated at varying concentrations of Ba(OH)2. AET was held constant at 100 mM and the base concentration varied. Results are shown in Figure 11.. When Ba(OH)2 was added to the reaction at a final concentration 50 mM or lower, little or no conversion was observed. Since one equivalent of barium binds two equivalents of chloride, and the AET used in these experiments was the hydrochloride, the effective concentration of Ba(OH)2 in the reaction was estimated to be 40 mM. Thus care must be taken to ensure that Ba(OH)2 is maintained in excess over AET in the reactions when method 1 is employed. As shown by the data of Figure 11,, the O-glycosylated peptide was converted under these conditions.
To examine the minimum amount of peptide that could be reacted with method 1, a mixture containing the 3.12-kDa tetraphosphopeptide (220 fmol), the 2.06-kDa singly phosphorylated peptide, and the glycosylpeptide was reacted in a volume of 10 μL. After 2 h at room temperature, 4 μL of 10% formic acid was added. The samples were desalted by reversed-phase as described in Materials and Methods and 1 μL was analyzed by MALDI-TOF-MS. Ignoring losses, the amounts loaded to the target were 44 fmol of 3.12-kDa tetraphosphopeptide, 135 fmol of the 2.06-kDa phosphopeptide, and 100 fmol of the glycosylated peptide (Fig. 22).). All peptides reacted to completion including the glycosylated peptide, which was unexpected, and in contrast to the earlier reports.7 In a parallel experiment, in which 28% NH4OH was substituted for Ba(OH)2, 60% of this glycopeptide remained unconverted even at the higher temperature of 45°C (data not shown)2.
A comparison of reaction conditions 0.09 M barium hydroxide at 25°C and 0.1 M sodium hydroxide at 45°C is shown in Figure 33.. The upper panel of the figure shows the reaction in barium hydroxide at 25°C with a half-life for the reacting peptide of about 15 min compared with 40 min for the same peptide in sodium hydroxide at 45°C (lower panel). Over the time course of the reaction, the singly protonated, tetraphosphorylated precursor ion fell to an undetectable level while the transient intermediates rose and fell. The data were consistent with four phosphorylated residues within the peptide and was consistent with no significant difference in chemical reactivity among the multiple sites. As a check, peak heights ratios were measured using a casein peptide contaminant as internal standard. Equivalent results were obtained (data not shown).13,14 The kinetic results were in general accord with those reported earlier for phosphopeptides.7
To avoid complications of counter ions, which react with the base, experiments were carried out in which β-elimination and addition reactions were performed separately. A mixture of the O-glycosylated and the two phosphopeptides was reacted in the same reaction vessel with Ba(OH)2 added to 45 mM, and β-elimination was allowed to proceed for 60 min at 25°C. AET was then added to 20 mM, and the addition reaction was carried out for another 60 min. As shown by the data of Figure 44,, all peptides were 100% converted. In a similar experiment, as little as 10 mM AET was required to give 100% formation of the aminoethylcysteine product. A control, which received no AET, was analyzed and showed only ions corresponding to the dehydroalanine products of all peptides (data not shown). No ions were observed that corresponded to serine residues formed by the back addition of water. This indicated that the unsaturated bond in dehydroalanine was stable under these experimental conditions and an O-glycosyl bond could be β-eliminated completely by Ba(OH)2.
Figure 55 shows the results of performing the chemistry on 4 pmol of the 3.12-kDa tetraphosphopeptide bound to a micro C18 ZipTip. As shown by the data, nearly complete conversion of all four sites was obtained in the 90-min reaction.
The product mixture from the 120-min time point from the experiment of Figure 33 (lower panel) was digested with Lys-C, desalted, and analyzed. Ions were recorded at 734.4, 1772.0, 2052.1, 2198.1, 2219.1, 2344.2, and 2365.2. These values were in good experimental agreement with the calculated singly charged ions for cleavage on the C-terminal side of serine residues 15, 17, 18, and 19. Results are summarized in the data of Table 22.
The CTID protocol was successfully conducted reacting less than 0.25 pmol of a 3-kDa tetraphosphopeptide from β-casein. β-Elimination of partially purified tryptic phosphopeptides was carried out in the presence of 0.1M AET and 0.09 M Ba(OH)2. The reaction went essentially to completion in 1 h at room temperature and no nonspecific modifications were detected. The use of the CTID protocol successfully assigned phosphoserine residues 15, 17, 18, 19, and 35 in β-casein. Reaction conditions for β-elimination were examined and it was determined that barium hydroxide at 25°C was preferred to 0.1 M NaOH at 45°C as reported by Byford.7 However, our results showed that O-glycosyl modification readily occurred and with kinetics essentially the same as for phosphopeptides. This result was in contrast to an earlier report with the protein mucin.7 Because of this result, caution is urged in relying on the ability of the method for differentiating glycosylated from phosphorylated residues.
By following the time course of the reaction, phosphorylation sites could be counted by using a semiquantitative method, which cast the multiple ion products as a relative ratio. This method may not be generally applicable to all phosphopeptides unless the products of the reaction possess sufficiently similar desorption efficiencies. The validity of the results given here and obtained with 2,5-dihydroxybenzoic acid matrix were confirmed using an internal, invariant peptide as reference ion (data not shown). Time-course data suggested the four sites within the tetraphosphopeptide displayed no preferential site reactivity regarding chemical conversion, and the proximity of acidic glutamic acid residues on the C- and N-terminal side of the modified serines did not interfere with cleavage by Lys-C or chemical conversion.
A dilution or a drying step has frequently been necessary when certain amines or thiols have been used as nucleophiles.6,7,11,12 The use of the hydrochloride salt of AET eliminated the need to add organic solvents to the reaction which obviated the requirement for its removal and thus facilitated desalting the reaction by reversed-phase.15 The absence of organic solvent also facilitated carrying out the reactions with peptide bound to reversed-phase resin; and, while silica-based resins normally tolerate very limited basic conditions, no breakdown in the ZipTip resin was observed. While the conditions used for reacting in the solid phase showed a high conversion of the four sites in the peptide, it should be noted that complete conversion was not necessary to identify target sites. Indeed, in selecting an intermediate reaction time, which included sites that remained phosphorylated, a fully interpretable mass spectrum was still obtained (Table 22).
It was convenient to use the hydrochloride of AET; however, its use required careful attention to the effects of the hydrochloride. Furthermore, sulfates, phosphates, and carbonates must be avoided when the base is barium. Results, which showed that peptides could be β-eliminated prior to nucleophilic addition, ameliorated the situation with respect to the hydrochloride salt and reduced the requirement for high concentrations of base. However, to avoid potential anionic interference, it is suggested that the chemistry be carried out in 90 mM Ba(OH)2, either with reversed-phase desalted peptides or with peptides bound to reversed-phase which is alkaline pH tolerant.
This work was supported in part by funds from the Beckman Institute at Caltech.