Increased efficiency in biomimetic Lewis acid–base pair catalyzed monoacylation of diols by acyl phosphate monoesters

The utility of acyl phosphates esters as biomimetic reagents for selective monoacylation of diols is a subject of continuing interest with important applications, including kinome interrogation (Patricelli et al. 2007; Nordin et al. 2015) and aminoacylation of the 3′-terminus of tRNA (Tzvetkova and Kluger 2007; Duffy and Dougherty 2010). We previously reported that lanthanide ions can serve as the chelated core to bind and activate the reacting species toward acyl transfer (Cameron et al. 2004; Tzvetkova and Kluger 2007; Her and Kluger 2011). Bidentate coordination of a cis-1,2-diol to a Lewis acid facilitates ionization of one of the coordinated diols while enhancing the activity of the bidentate-coordinate acyl phosphate (Scheme 1). The bis-bidentate array of coordinated Lewis base (diol) and Lewis acid (acyl phosphate) promotes the ultimate combination that produces the acylated diol along with a phosphate monoester as a by-product. The arrangement can utilize the potential energy made available by the separation of the Lewis acid–base pair. Utilizing acyl phosphate monoesters as metal-coordinated acyl donors at the same Lewis acid where the ionized diol is coordinated promotes the reaction without involvement of the solvent. With a similar approach, regioselective monoacylation of carbohydrates in water can be based on the geometry of adjacent hydroxyl groups (Gray and Kluger 2007; Dhiman and Kluger 2010). However, although the system is set up for efficient acylation, the formation of the correct combination of ligands in the chelate for reaction is subject to competing homologous formation of the chelates as well as interference due to coordination to the lanthanide of the phosphate monoester product (see below).

An important potential use of the approach is for biomimetic aminoacylation of tRNA for the ribosomal introduction of unnatural amino acids into new proteins, which requires an aminoacylated reagent whose reactive functionality parallels that of the biological agents, aminoacyl adenylates (Schimmel 1987). However, complications arise from the competing effectiveness of the lanthanum ion in the hydrolysis of phosphate derivatives that are the acylation reagents (Hendry and Sargeson 1989; Kluger et al. 1997). Furthermore, transfer of the acyl group, whether to the hydroxyl or to water, releases ethyl phosphate, forming an insoluble complex with the metal ions that prevents further catalysis (Tzvetkova 2008).

In the present study, we assessed the effects of other metal ions for reactions in water that are reported to be effective in nonaqueous systems (Hikawa et al. 2014). We also assessed aqueous reaction conditions with added solvents and at lower reaction temperatures. In some cases, these provided significant improvements for the overall process.

An aminoacyl phosphate monoester, PheEtP, was used as the acylation agent for the evaluation of catalysts in this study. This reagent is a good model for the general approach to the formation of amino acids esters at the 3′-terminal of tRNA. We found that conducting reactions in mixtures of DMSO and water reduced the rate of hydrolysis of PheEtP in the presence of lanthanum ion compared with the reaction in water alone. In a 95/5 DMSO/aqueous buffer (v/v), the t _½ of PheEtP was about 30 min, whereas the t _½ in water was <2 min. Nonetheless, the yield of acylated products did not increase with the reduced hydrolysis rate of the reagent. Surprisingly, the yield of ester derivatives that were formed at the 2′- or 3′-ribosyl positions of AMP with PheEtP remained near 30% after 24 h, regardless of the amount of DMSO in the aqueous reaction mixture. This suggests that the reactants are locally solvated exclusively by water. The polarity of the reaction partners and their ability to form hydrogen bonds is consistent with this outcome. It is likely that the ionization of water coordinated to lanthanum ions is essential for the activation of PheEtP, whereas the reactivity of the hydroxyl group of diol toward the carboxyl carbon is increased as it is deprotonated by the lanthanum-coordinated hydroxyl (Kluger and Cameron 2002). Lowering the relative amount of water reduces the extent of ionization of the hydroxyl of the diol moiety, reducing the reactivity in the coordinated ligands.

The problematic competing hydrolysis of PheEtP occurs by the addition of non-chelated water. This will have a higher enthalpic barrier if it avoids the coordination sphere of the metal ion. Lowering the reaction temperature should reduce the rate of the reaction with water to a greater extent than it will reduce the desired acylation from reaction with the bis-bidentate coordinated chelate. As expected, at 4 °C, hydrolysis occurred at a significantly slower rate than at room temperature. At the lower temperature, 15% of the initial PheEtP remained after 2.5 h, corresponding to a t _½ of 1 h under these conditions. Applying the same conditions to the lanthanum ion-promoted aminoacylation of AMP increased the yield from around 30% to 40%. For aminoacylation of dCA, a dinucleotide used by Robertson et al. (1989) in the chemical aminoacylation of tRNA, the combined yields of 2′ and 3′ esters doubled from 6.8% to 15% after 12 h of reaction. With PheEtP at its saturating concentration, the reaction produced the desired esters at a 40% yield after 12 h at 4 °C (Table 1). This establishes the importance of using lower reaction temperatures, consistent with the expected effect of selectively slowing the competing hydrolysis.

With this useful outcome, we optimized conditions for acylation with metal salts and determined the extent of the competing hydrolysis of PheEtP. The t _½ for catalyzed hydrolysis of PheEtP at 4 °C in the presence of zinc or lead salts were 60 and 5 min, respectively. The reaction in the presence of cupric salts is not effective, as it selectively promotes the hydrolysis of PheEtP. Based on these observations, we anticipated that Zn²⁺ and Pb²⁺ should catalyze aminoacylation of diols at or near 0 °C. We tested this approach by following the aminoacylation of ethylene glycol by PheEtP at 4 °C (Table 2).

In the presence of Zn²⁺ or Pb²⁺, the reactions of ethylene glycol and PheEtP produced a mixture of the ester and the hydrolysis product, phenylalanine, within 1 min after addition to the solution. The Pb²⁺-catalyzed reactions occurred with a half-life for PheEtP of <1 min. The yields are comparable with those from reactions conducted in the presence of La³⁺. The Zn²⁺-catalyzed reactions were slower and their overall yields were lower than with Pb²⁺ and La³⁺.

We also investigated the reactions of ethylene glycol with BMP (Scheme 2). The t _½ for hydrolysis of BMP in the presence of Zn²⁺ and Pb²⁺ were 24 and 1 h, respectively. The Pb²⁺-catalyzed monobenzoylation reactions of ethylene glycol gave the best yield of the monoester after 3 h, whereas reactions with Zn²⁺ required about 12 h to reach their maximum yield.

We also evaluated the effectiveness of Zn²⁺ and Pb²⁺ as catalysts for reactions of PheEtP with a mononucleotide and a dinucleotide. Both AMP and dCA were separately incubated with PheEtP in the presence of the metal ions at 4 °C, followed by analysis that used reverse-phase HPLC to separate the reactants and products. The reactions gave 2′ and 3′ nucleotidyl esters in each case. In reactions with dCA, the yield was 19% with Zn²⁺ and 13% with Pb²⁺ (Figures 1 and 2). These did not improve on the previously reported 15% yield with La³⁺. We also monitored the reactions for an additional 72 h during which the distribution of products remained unchanged, establishing that the esters are stable under these conditions and that the reactions are self-limiting.

Reports of previous studies indicate that Pb²⁺ is among the most effective catalysts for site-specific cleavage of RNA (Huff et al. 1964; Brown et al. 1983). Lead hydroxide is an active catalyst for RNA hydrolysis (Brown et al. 1983) and for transesterification of a phosphate diester (Morrow et al. 1992). In reactions at pH around 7, where Pb²⁺-promoted aminoacylation occurs, lead hydroxides are present (Perera et al. 2001). Lead hydroxides could assist in deprotonation of the incoming hydroxyl, further promoting acylation.

As noted earlier, a major problem with La³⁺-promoted acylation is the high affinity of La³⁺ to phosphate monoesters. The resulting precipitate removes the La³⁺ catalyst from solution. As a result, La³⁺-promoted acylation reactions require a large excess of the added metal ion (Gray and Kluger 2007). Reactions with Zn²⁺ and Pb²⁺ reduce the formation of insoluble metal phosphates, providing alternative catalysts for acylation reactions via acyl phosphates that can operate at lower concentrations.

We have demonstrated that lowering the reaction temperature improves lanthanide-ion-promoted aminoacylation reactions by selectively slowing the competing hydrolysis of activated amino acyl phosphates. Catalysis by Zn²⁺ and Pb²⁺ is effective, producing the desired aminoacylation products under conditions where the acyl phosphate reagent has improved stability.

We thank NSERC Canada for support through a Discovery Grant.