|Home | About | Journals | Submit | Contact Us | Français|
Protein lysine methylation is a biologically-important posttranslational modification. Given the increased use of alkyllysine-containing peptides to study protein methylation, we described an efficient solid-phase synthesis to access various alkyllysine-containing peptides. As an exemplary application, we evaluated how ionization efficiency of the alkyllysine peptides affects mass spectrometric analysis.
Lysine-Nε amino groups of proteins undergo diverse posttranslational modifications such as ubiquitination, sumolyation, neddylation, acetylation and methylation.1 Among the most known lysine posttranslational modifications is methylation, which is featured by enzymatic installation of a methyl group from the biological methyl donor S-adenosyl-L-methionine (SAM) (Fig. 1).2 Despite its stable chemical nature, this modification can be removed by demethylases (Fig. 1).2 The dynamic methylation and demethylation of targeted proteins play essential roles in epigenetic regulation via modulating their localization, stability or interacting partners.3 Their deregulation has been linked to many diseases including cancer.3 For instance, histone H3 lysine 9 (H3K9) can be methylated by 7 human protein lysine methyltransferases (PKMTs).4 This modification is associated with gene silencing at euchromatin, facultative/constitutive heterochromatin, or virus-intergrating sites to maintain normal cellular functions.5, 6 In contrast, the same events can be hijacked by cancer cells to silence tumor suppressor genes.7
Given the importance of protein methylation, multiple approaches were described for site-specific incorporation of methylated lysines into peptides or proteins.8–10 Complementary biochemical assays were also developed to trace these events.10 Recently, the peptides that contain lysine-Nε-β unsaturated substituents have emerged as valuable chemical entities to study protein methylation and demethylation.11–15 For example, several successes have been reported to utilize SAM derivatives, in combination with native or engineered PKMTs, to transfer allyl derivatives to PKMT substrates (Fig. 1).13–15 The resultant lysine-Nε derivatization enabled the PKMT substrates to be readily characterized.13, 14 The peptides carrying lysine-Nε-propargyl/allyl derivatives are potent inhibitors of the protein lysine demethylase LSD1 (Fig. 1).11, 12 These applications thus triggered us to explore efficient chemical methods to assemble Nε-β-unsaturated moieties onto lysine side chain.11, 12
Among known synthetic strategies to access Nε-alkyllysines are base-promoted lysine-Nε alkylation, lysine-Nε reductive alkylation, and leaving-group-facilitated SN2 C-amination of 6-hydroxynorleucine.11, 12, 16, 17 Base-promoted lysine-Nε-alkylation is less selective for the degree of alkylation and often leads to mono/di/tri-alkylated product mixtures.8, 9 Lysine-Nε reductive alkylation is facile to prepare Nε-monomethyllysine. The monoalkylation selectivity was achieved by reductive benzylation and then methylation through imine intermediates followed by debenzylation.8, 9 However, this approach is not suitable to install lysine-Nε-β-unsaturated substituents because of the undesirable loss of the unsaturated moieties via 1,4-hydride reduction of the imine intermediates.c So far, the reported approach to chemically install lysine-Nε-β-sp/sp2 substituents is to mesylate 6-hydroxynorleucine of resin-immobilized peptides, followed by solution-phase SN2 displacement of the mesylate with Nβ-unsaturated amines (SN2-type C-amination), in which multiple HPLC purifications were involved.11, 12
Here we described a concise, efficient chemical approach to install lysine-Nε-β-unsaturated substituents via solid-phase Dess-Martin oxidation/reductive amination of hydroxynorleucine. Characterizing these lysine-Nε-derivatized peptides also allowed us to compare ionization efficiency of diverse Nε-alkyllysine-containing N-terminal H3 peptides on matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). Although most H3K9-Nε alkyl substitutions (e.g. Me-, CF3CH2-, propargly, allyl, 2-butynly and 2-butenyl) have no significant effect on the ionization efficiency, ethylation and benzylation alter the ionization efficiency by 2.5–5-fold. These results therefore claimed the need to validate ionization efficiency of native and alkyllysine peptides for direct quantification through their MALDI-MS peak intensities.
To develop strategies to access lysine derivatives containing Nε-β-unsaturated substituents, we first focused on solution-phase synthesis of Nε-allyllysine as a proof-of-concept target (Scheme 1). The reductive alkylation reaction of Fmoc-protected lysine and acrolein underwent undesirable 1,4– and then 1,2–imine reduction to give Nε-propyl rather than Nε-allyllysine.c To avoid this 1,4–hydride reduction, we envisioned a strategy through switching the positions of the amine-aldehyde pair (reductive alkylation versus reductive amination). Here Nα-Fmoc lysine was converted to Nα-Fmoc hydroxynorleucine 1 via spontaneous hydrolysis of a diazonium intermediate (Scheme 1).11, 12, 18 A subsequent direct oxidation of hydroxynorleucine’s primary alcohol to aldehyde proved to be challenging. All the attempts for such oxidation, concomitant with the loss of 1, gave a mixture of products, which either showed complex profiles (e.g. with tetra-n-propyl ammonium perruthenate, pyridinium chlorochromate, or chromium(VI) oxide/pyridine as oxidants) or decomposed in the course of purification (e.g. oxalyl chloride/DMSO for Swern oxidation or periodinane for Dess-Martin oxidation).
Given that the product profile under the Dess-Martin condition is simpler than others, we reasoned that the failure of Dess-Martin periodinane to yield the desirable aldehyde could be due to the challenge of product purification. This problem was gratifyingly circumvented using a step-wise one-pot reaction (Scheme 1). Here Nα-Fmoc hydroxynorleucine was first oxidized to the corresponding aldehyde 2 with Dess-Martin periodinane. After simple concentration, the aldehyde product was reacted with allyl amine (2 equiv. to suppress multi-alkylation), followed by cyanoborohydride-mediated reduction, to yield Nε-allyllysine 3. Without further purification, Nε-allyllysine was masked by Boc-carbamate with a final yield of 15% from Nα-Fmoc hydroxynorleucine 1 to 4a. The generality of this synthesis was also confirmed by its application to prepare Nα-Fmoc-Nε-(Boc, propargyl)-lysine 4b and Nα-Fmoc-Nε-(Boc, 2-butynyl)-lysine 4c in comparable 15–20% yields (Scheme 1, Section 2b in ESI). Despite less promising yields, the one-pot strategy served as a key starting point to develop more efficient solid-phase synthesis (results below).
Encouraged by the success of solution-phase synthesis of Nε-allyl/propargyl/2-butynyl lysine analogues 4a–c, we then explored such chemical transformation of a peptide substrate in solid phase (Scheme 2). The N-terminal H3 peptide (3–13 aa, 5) with its H3K9 replaced by hydroxynorleucine was assembled on Wang resin. In comparison with the prior solid-phase peptide synthesis (SPPS) for hydroxynorleucine-containing peptides,11, 12 the current approach is featured by incorporating hydroxynorleucine and other amino acids without protecting hydroxynorleucine ε-hydroxyl group. Similar to our solution-phase strategy, the hydroxynorleucine moiety of resin-immobilized peptide was efficiently converted to aldehyde by Dess-Martin periodinane (from 5 to 6). Subsequent solid-phase reductive amination of the aldehyde with 2 equiv. ally amine and then TFA-mediated global deprotection yielded the desirable allyl peptide 8a with a remarkable yield of 30% for 23-step solid-phase reactions from Wang resin. The same approach also allowed the synthesis of 7 other H3K9-alkylated peptides (Me-, ethyl, CF3CH2-, propargly, allyl, 2-butynly, 2-butenyl and benzyl) from respective primary amines (30–50% yields from 1 to 8a–h based on Wang Resin, Scheme 2, Fig. S2 and Section 2c in ESI). The current strategy is convenient because of its ability to access multiple products from the commonly-shared hydroxynorleucine peptide via simple solid-phase washing and a single HPLC purification. In addition, Thr(tBut), Lys(Boc), Gln(Trt), Arg(Pbf), Ser(tBut), Gly and Cys(Trt) are tolerant to Dess-Martin oxidation/reductive amination conditions.d The robust and general applicability of the solid-phase Dess-Martin oxidation/reductive amination thus presents a short and efficient route to accessing various alkyllysine-containing peptides.
To examine the activities of protein methyltransferases, a common strategy is to quantify enzymatic modification with MALDI-MS. Some prior experiments relied on direct comparison of ion intensities of modified peptides (products) versus unmodified peptides (starting materials).19 However, given uncertain ionization efficiency of the alkylated peptides, the caution should be made for such direct comparison.13, 15, 20 For example, prior to MALDI-based MS quantification, comparable ionization efficiency on MALDI-MS was confirmed for N-terminal H3 peptide and its (E)-hex-2-en-5-ynyl derivative.13
Given our newly-gained ability to access various alkyllysine peptides, their ionization efficiency on MALDI-MS were evaluated in a systematic manner. Upon co-injecting equal-molar N-terminal H3 3–13 aa peptide and its H3K9-alkylated derivatives (Me-, Et-, CF3CH2-, propargly, allyl, 2-butynly, 2-butenyl and benzyl) on MALDI-MS, their ion intensities were compared (Figs. 2, S2 in ESI). Although the H3K9 peptides containing Me-, CF3CH2-, propargly, allyl, 2-butynly and 2-butenyl modifications showed MALDI-MS ionization efficiency comparable to that of the native peptide, the ethylation reduces and the benzylation enhances the ionization efficiency by 2.5- and 5-fold, respectively (Fig. 2). This finding thus argues that, although H3K9 methylation, allylation and propargylation have negligible effect on the peptide’s MALDI-MS ionization efficiency, other H3K9 alkylations (e.g. ethylation or benzylation) may alter such parameter dramatically.
In conclusion, a convenient synthesis has been developed through consecutive solid-phase Dess-Martin oxidation and reductive amination. The unprecedented approach enables the ready access to alkyllysine-containing peptides via SPPS, in particular those containing lysine-Nε-β-allyl/propargyl analogues. Given that similar peptide derivatives have been explored as intermediates for labelling PKMT substrates and as inhibitors of histone-modifying enzymes, these applications can be benefited by diversifying the peptide entities with the delineated synthesis. Analyzing ionization efficiency of alkyllysine-containing peptides also underscores the importance of validating the MS parameter upon applying MALDI-MS to quantify alkylated peptides. Here Nε-alkyllysine-containing peptides can be readily prepared as the MS standards for such validation through the described synthesis.
We thank Dr. Ouathek Ouerfelli for helpful discussion, Drs. George Sukenick, Hui Liu, Hui Fang for mass spectral analysis, and the financial supports from NIH through NIGMS (1R01GM096056-01) and the NIH Director's New Innovator Award Program (1-DP2-OD007335-01), March of Dimes Foundation through Basil O’connor Starter Scholar Award, The Starr Cancer Consortium, Alfred W. Bressler Scholars Endowment Fund, Mr. William H. Goodwin and Mrs. Alice Goodwin the Commonwealth Foundation for Cancer Research and The Experimental Therapeutics Center of Memorial Sloan-Kettering Cancer Center. D.C is supported by the graduate fellowship of Tri-Institutional Training Program in Chemical Biology.
†Electronic Supplementary Information available: General materials & methods, synthetic protocols and product characterization. See DOI: 10.1039/b000000x/
cUpon reacting Na-Fmoc-lysine with acrolein, NaCNBH3 and then Boc2O, Na-Fmoc-Nε-(Boc, propyl)lysine was isolated as the only major product, which arises from 1,4–imine reduction (Section 2a of ESI).
dThr(tBut), Lys(Boc), Gln(Trityl), Arg(Pbf), Ser(tBut) and Gly were installed in the course to synthesizing the 3–13 aa H3 peptides. The success of the peptide synthesis therefore indicated that these amino-acid building blocks are tolerant to Dess-Martin oxidation/reductive amination. The tolerance of Cys(Trt) was confirmed independently by synthesizing C-terminal-Cys(Trt)-apppended 3–13 H3 peptide (data not shown). Other amino acids (Met, Tyr and Trp) remained to be examined.